From Argument Schemes To Argumentative Relations In The Wild: A Variety Of Contributions To Argumentation Theory 3030283666, 9783030283667, 9783030283674

Table of contents :
Contents......Page 6
Contributors......Page 8
1 A Variety of Contributions to Argumentation Theory......Page 10
References......Page 18
2.1 The Notion of Argument Scheme......Page 20
2.2 Intersubjective Procedures for Evaluating Argumentation......Page 22
2.3 The Pragma-Dialectical Typology of Argument Schemes......Page 23
2.4 Critical Questions: Pragmatic Argumentation as a Case in Point......Page 26
2.5 Subtypes and Variants of Argumentation......Page 28
References......Page 31
3 In Search of a Workable Auxiliary Condition for Authority Arguments......Page 33
3.1 Corroboration......Page 34
3.2 Trust......Page 37
3.3 The Layperson’s Paradox......Page 41
3.4 Goldberg Variations......Page 43
3.5 Discussion......Page 46
References......Page 48
4.1 Introduction......Page 49
4.2.1 Normative Issues......Page 50
4.2.2 Salvaging the Standard Treatment: Rhetorical Insights......Page 51
4.3.1 Cognitive Underpinnings of Argument Evaluation......Page 53
4.3.2 Undetected Fallacies Are Arguments That Are Non-manifest Qua Fallacies......Page 56
4.3.3 Pragma-Linguistic Constraints on Argumentation and Rhetorical Effectiveness......Page 58
4.4 A Fallacious and Deceptive Erotetic Tour de Force......Page 61
4.4.1 A Double Rhetorical Trick Question Realising a False Dilemma......Page 62
4.4.2 More Rhetorical Questions to Strengthen the Context......Page 64
4.4.3 The Straw Man Question......Page 65
4.5 Conclusion......Page 68
References......Page 69
5.1 Introduction......Page 71
5.2 Argumentation in Philosophy: The Problems of Evaluation Standars......Page 73
5.3.1 What Is the Genetic Fallacy?......Page 74
5.3.2 The Genetic Fallacy in Question......Page 77
5.4 The Charge of Committing a Fallacy as Strategic Maneuvering......Page 78
5.5 Conclusion......Page 82
References......Page 83
6.1 The Role of Questions in Argument......Page 86
6.2 Burden of Proof......Page 88
6.3 Typical Moves in Philosophical Argumentation......Page 90
6.4 Revealing Irrelevant Distinctions......Page 91
6.5 Analytic Dilemmas......Page 97
6.6 Pragmatic Self-refutation......Page 99
6.7 Conclusions......Page 100
References......Page 102
7 Eudaimonistic Argumentation......Page 104
7.1 Virtue Theories of Argumentation......Page 105
7.2 Intellectual Flourishing......Page 108
7.3 Adversariality......Page 110
References......Page 112
8.1 Introduction......Page 114
8.2 The Importance of Community Argument......Page 115
8.3 Inherent Barriers to Community Argument......Page 116
8.3.1 Evolved Sociality......Page 117
8.3.2 Evolved Cognitive Modularity......Page 120
8.4.1 Distraction......Page 123
8.4.2 Nudges......Page 125
8.4.3 Narratives......Page 126
8.4.4 Visual Persuasion......Page 128
8.4.5 Some New Horizons......Page 130
8.5 Community Argument? Concerns and Ambitions......Page 131
References......Page 134
9.1 Warrants, Sources of Backing, and Rules......Page 138
9.2 From Bodies of Law to Formally Backed Warrants......Page 139
9.3 Reliably Understood Transcripts and Speech Acts......Page 141
9.4 Reliable Institutional Warrants, Rebuttals, and Argument Cogency......Page 144
9.5 Recognizing Rebuttals......Page 145
9.6 Precedents, Conclusive Versus Defeasible Warrants, and Refining Warrants......Page 147
9.7 Defeasible Warrants and Warrant Strength......Page 149
9.8 Warrant Strength and Argument Cogency......Page 151
9.9 Application to Institutional Warrants in General......Page 152
References......Page 155
10.1 Introduction......Page 156
10.3 Dismissal......Page 157
10.4 Objection......Page 158
10.5 Rebuttal......Page 159
10.6 Refutation......Page 162
10.7 Weighing......Page 164
10.8 Rebuttals, Refutations and Standards of Proof......Page 165
10.9 Counterargument Structures......Page 166
10.10 Other Classifications......Page 168
10.11 The Order of Counterargumentation......Page 170
10.12 Conclusions......Page 171
References......Page 172
11.1 Introduction......Page 174
11.2 Arguments for Questions?......Page 175
11.3 The Formal Logic of Inferences to Questions......Page 177
11.4.1 Why Questions......Page 180
11.4.2 How Questions......Page 182
11.4.3 What Questions......Page 183
11.5 Inferences from Questions to Questions......Page 186
11.6 Summary......Page 189
References......Page 190
12.1 Introduction......Page 192
12.2.1 Apprehensive Straks as a Construction......Page 194
12.2.2 Argumentative Semantics......Page 198
12.3 Apprehensive Straks in Argumentative Discourse......Page 199
12.3.1 Pragmatic Argumentation: Warnings and Slippery Slope......Page 200
12.3.2 Reductio ad Absurdum......Page 202
12.4 Expressives Between Reasonableness and Effectiveness......Page 203
12.4.1 The Expressive Dimension......Page 204
12.4.2 Overt Unreasonableness......Page 206
12.5 Conclusion......Page 208
References......Page 209
13.1 Introduction......Page 212
13.2 Abduction......Page 213
13.3 Conversational Implicature and Irony......Page 216
13.4.1 Irony: Standpoint or Conclusion?......Page 220
13.4.2 Irony as Strategic Maneuvering......Page 222
13.4.3 Interpreting Irony by Argumentative Abduction......Page 223
13.5 Conclusion......Page 226
References......Page 227
14.1 Introduction......Page 229
14.2 The Current Consensus: Representing “Conductive Argument” as a Single Argument......Page 232
14.3 My Proposal: Representing Pro/Con Argumentation as Deliberative Process......Page 233
14.4 Conclusion......Page 241
References......Page 242
15.1 Introduction......Page 244
15.2.1 The Supplementation of On-balance Premise Approach......Page 245
15.2.2 The Warrant-Reformulation Approach......Page 247
15.3 The Linguistic Feature of Conductive Arguments......Page 248
15.4 The Metaphor of Outweighing......Page 251
15.5 The Role of Counter-Considerations in Conductive Arguments......Page 253
15.6 Conclusion......Page 257
References......Page 258
16.1 Introduction......Page 260
16.2 “Adler’s Problem” and Its Variant......Page 262
16.3 The Rhetorical Solution Versus the Logical Solution......Page 265
16.4 The Perspective of Argument Evaluation......Page 268
16.5 Conclusion......Page 271
References......Page 272
17.1 Introduction......Page 273
17.2 Related Work......Page 274
17.3.1 Training Dataset......Page 275
17.4.1 Dialogue Excerpt......Page 276
17.4.2 Annotation Results......Page 278
17.4.3 Mining Argumentative Relations “in the wild”......Page 281
17.5.1 Merged Arguments......Page 282
17.6 Conclusion......Page 285
References......Page 288
Index......Page 290

Citation preview

Argumentation Library

Frans H. van Eemeren Bart Garssen   Editors

From Argument Schemes to Argumentative Relations in the Wild A Variety of Contributions to Argumentation Theory

Argumentation Library Volume 35

Series Editor Frans H. van Eemeren, University of Amsterdam, Amsterdam, The Netherlands Editorial Board Maurice A. Finocchiaro, University of Nevada, USA Bart Garssen, University of Amsterdam, The Netherlands David Zarefsky, Northwestern University, USA

Since 1986 Springer, formerly Kluwer Academic Publishers, publishes the international interdisciplinary journal Argumentation. This journal is a medium for distributing contributions to the study of argumentation from all schools of thought. From a journal that published guest-edited issues devoted to specific themes, Argumentation has developed into a regular journal providing a platform for discussing all theoretical aspects of argumentative discourse. Since 1999 the journal has an accompanying book series consisting of volumes containing substantial contributions to the study of argumentation. The Argumentation Library aims to be a high quality book series consisting of monographs and edited volumes. It publishes texts offering important theoretical insights in certain major characteristics of argumentative discourse in order to inform the international community of argumentation theorists of recent developments in the field. The insights concerned may pertain to the process of argumentation but also to aspects of argumentative texts resulting from this process. This means that books will be published not only on various types of argumentative procedures, but also on the features of enthymematic argumentation, argumentation structures, argumentation schemes and fallacies. Contributions to the series can be made by scholars from a broad variety of disciplines, ranging from law to history, from linguistics to theology, and from science to sociology. In particular, contributions are invited from argumentation theorists with a background in informal or formal logic, modern or classical rhetoric, and discourse analysis or speech communication. A prerequisite in all cases is that the contribution involved is original and provides the forum of argumentation theorists with an exemplary specimen of advanced scholarship. The Argumentation Library should enrich the study of argumentation with insights that enhance its quality and constitute a fruitful starting point for further research and application. All proposals will be carefully taken into consideration by the editors. They are to be submitted in fourfold. If the prospects for including a certain project in the series are realistic, the author(s) will be invited to send at least three representative chapters of their manuscript for review to the editors. In case the manuscript is then judged eligible for publication, the complete manuscript will be reviewed by outside expert referees. Only then a final decision can be taken concerning publication. Authors interested in submitting a proposal or completed manuscript can contact either [email protected] or the Series Editor.

More information about this series at http://www.springer.com/series/5642

Frans H. van Eemeren Bart Garssen •

Editors

From Argument Schemes to Argumentative Relations in the Wild A Variety of Contributions to Argumentation Theory

123

Editors Frans H. van Eemeren University of Amsterdam Amsterdam, The Netherlands

Bart Garssen University of Amsterdam Amsterdam, The Netherlands

ISSN 1566-7650 ISSN 2215-1907 (electronic) Argumentation Library ISBN 978-3-030-28366-7 ISBN 978-3-030-28367-4 (eBook) https://doi.org/10.1007/978-3-030-28367-4 © Springer Nature Switzerland AG 2020 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Contents

1

1

A Variety of Contributions to Argumentation Theory . . . . . . . . . . Frans H. van Eemeren and Bart Garssen

2

Argument Schemes: Extending the Pragma-Dialectical Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Frans H. van Eemeren and Bart Garssen

11

In Search of a Workable Auxiliary Condition for Authority Arguments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hans V. Hansen

25

Give the Standard Treatment of Fallacies a Chance! Cognitive and Rhetorical Insights into Fallacy Processing . . . . . . . . . . . . . . . Steve Oswald and Thierry Herman

41

Argument Evaluation in Philosophy: Fallacies as Strategic Maneuvering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Federico E. López

63

Dialogical Sequences, Argumentative Moves and Interrogative Burden of Proof in Philosophical Argumentation . . . . . . . . . . . . . . Joaquin Galindo

79

3

4

5

6

97

7

Eudaimonistic Argumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Andrew Aberdein

8

Worries About the Prospects for Community Argument . . . . . . . . 107 Dale Hample

9

Assessing Connection Adequacy for Arguments with Institutional Warrants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 James B. Freeman

v

vi

Contents

10 On the Logical Ways to Counter an Argument: A Typology and Some Theoretical Consequences . . . . . . . . . . . . . . . . . . . . . . . . 149 Hubert Marraud 11 Arguing for Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 David Hitchcock 12 Expressives in Argumentation: The Case of Apprehensive Straks (‘Shortly’) in Dutch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 Ronny Boogaart 13 Argumentative Abduction in the Interpretation Process: A Pragma-Dialectical Study of an Ironic Utterance . . . . . . . . . . . . 205 Antonio Duarte 14 Is “Conductive Argument” a Single Argument? . . . . . . . . . . . . . . . 223 Isabela Fairclough 15 On the Logical Reconstruction of Conductive Arguments . . . . . . . . 239 Yun Xie 16 The Legitimacy of Conductive Arguments: What Are the Logical Roles of Negative Considerations? . . . . . . . . . . . . . . . . . . . . . . . . . 255 Yanlin Liao 17 Deploying Machine Learning Classifiers for Argumentative Relations “in the Wild” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269 Oana Cocarascu and Francesca Toni Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287

Contributors

Andrew Aberdein School of Arts and Communication, Florida Institute of Technology, Melbourne, USA Ronny Boogaart Leiden University Centre for Linguistics, Leiden, The Netherlands Oana Cocarascu Department of Computing, Imperial College London, London, UK Antonio Duarte Department of Logic and Theoretical Philosophy, Universidad Complutense de Madrid, Madrid, Spain Isabela Fairclough University of Central Lancashire, Preston, UK James B. Freeman Department of Philosophy, Hunter College of the City University of New York, New York, USA Joaquin Galindo Department of Philosophy, University of Valladolid, Madrid, Spain Bart Garssen Speech Communication, Argumentation Theory and Rhetoric, ILIAS, University of Amsterdam, Amsterdam, The Netherlands Dale Hample Department of Communication, University of Maryland, College Park, USA Hans V. Hansen Department of Philosophy, University of Windsor, Ontario, Canada Thierry Herman Institute of Information and Language Science, University of Neuchâtel, Neuchâtel, Switzerland; School of French as a Foreign Language, University of Lausanne, Lausanne, Switzerland David Hitchcock McMaster University, Hamilton, Canada Yanlin Liao Modern Logic and Application Institute/Department of Philosophy, Nanjing University, Nanjing, Jiangsu, China vii

viii

Contributors

Federico E. López Centro de Investigaciones Filosóficas, Facultad de Humanidades y Ciencias de la Educación, Universidad Nacional de La Plata, Ensenada, Buenos Aires, Argentina Hubert Marraud Department of Linguistics, Modern Languages, Logic and Philosophy of Science, Theory of Literature and Comparative Literature, School of Arts, Universidad Autónoma de Madrid, Madrid, Spain Steve Oswald English Department, University of Fribourg, Fribourg, Switzerland Francesca Toni Department of Computing, Imperial College London, London, UK Frans H. van Eemeren Speech Communication, Argumentation Theory and Rhetoric, ILIAS, University of Amsterdam, Amsterdam, The Netherlands; Leiden University, Leiden, The Netherlands Yun Xie Department of Philosophy, Institute of Logic and Cognition, Sun Yat-sen University, Guangzhou, China

Chapter 1

A Variety of Contributions to Argumentation Theory Frans H. van Eemeren and Bart Garssen

In From argument schemes to argumentative relations in the wild we have brought together a variety of contributions to argumentation theory that we want to bring to the attention of the international community of argumentation scholars and students of argumentation. The chapters collected in this volume are all based on presentations given at the 9th Conference of the International Society for the Study of Argumentation (ISSA) held in Amsterdam at the beginning of July 2018. They cover a broad spectrum of theoretical issues, ranging—as the title of this volume indicates—from the definition and classification of argument schemes to the treatment of argumentative relations “in the wild”. Although the various chapters are written from different theoretical perspectives, which are not always compatible with each other, each of them constitutes in our view a worthwhile contribution to the necessary reflection upon the problems involved in the theorizing about argumentation. The first chapters of this volume concentrate on the theoretical concepts of argument schemes and fallacies, starting with a pragma-dialectical view on how to develop a theoretical approach to argument schemes satisfying the principle of Occam’s razor that a simpler solution to a problem is to be favoured to more complex solutions, so that a simpler classification of argument schemes is in principle to be preferred to a more complex one. Next some problems involved in dealing with a specific kind of argument scheme, i.e. the one employed in authority argumentation, are discussed. The second theoretical issue given its due in the first part of the volume, the treatment of the fallacies, is discussed in two other contributions. In one of them a proposal F. H. van Eemeren (B) · B. Garssen Speech Communication, Argumentation Theory and Rhetoric, ILIAS, University of Amsterdam, Amsterdam, The Netherlands e-mail: [email protected] B. Garssen e-mail: [email protected] F. H. van Eemeren Leiden University, Leiden, The Netherlands © Springer Nature Switzerland AG 2020 F. H. van Eemeren and B. Garssen (eds.), From Argument Schemes to Argumentative Relations in the Wild, Argumentation Library 35, https://doi.org/10.1007/978-3-030-28367-4_1

1

2

F. H. van Eemeren and B. Garssen

is made to give the “standard treatment” of the fallacies another chance by taking cognitive and rhetorical insights into fallacy processing into account. In another contribution, focusing on the “genetic fallacy”, argument evaluation in philosophy is examined from a different perspective by viewing such a fallacy pragma-dialectically as a derailment of strategic manoeuvring. In ‘Argument schemes’ (Chap. 2) Frans van Eemeren and Bart Garssen turn to the theoretical issue of how in argumentation an effort is made to transfer acceptance of the reasons that are advanced in support of a standpoint to the standpoint defended. They explain their approach by recapitulating the rationale of the pragma-dialectical perspective on argument schemes and extending its general outlines. In the pragmadialectical view (van Eemeren & Grootendorst 1992: 94–102), argument schemes are schematic representations of justificatory relationships between reasons advanced in support of a standpoint and the standpoint defended that are instrumental in legitimizing a transfer of acceptability from these reasons to the standpoint concerned. In Chap. 2 van Eemeren and Garssen discuss the pragmatic and dialectical rationales for distinguishing between the various argument schemes utilized in argumentative discourse. Each of these argument schemes manifests itself in the discourse in specific types and subtypes of argumentation. Which types and subtypes of argumentation can be used in resolving a difference of opinion and how their use is to be evaluated is in the pragma-dialectical view to be agreed upon in the opening stage of a critical discussion when the parties involved determine their joint starting points. In actual argumentative practices such an agreement will in a great many cases already be presupposed. ‘In search of a workable auxiliary condition for authority arguments’ (Chap. 3) by Hans Hansen focuses on an auxiliary condition applying to the use of an argument scheme for appeal-to-authority arguments. The purpose of this condition is to give laypersons an additional reason to accept the speaker’s claim over and above the speaker’s claim to expertise. In Chap. 3 Hansen explores three different ways of satisfying the auxiliary condition. In two of them auxiliary experts play a part. One of them casts these experts as corroborators of what the cited authority has asserted; in the other one they are endorsers of the epistemic character of the source. In both cases laypeople who want to use an authority argument meet with practical difficulties. Hansen also considers a third auxiliary condition, based on an idea advanced by Goldberg (2011), according to which laypersons must engage with the informational sources accessible to them and seek out credible objections to what the main source has proposed. If they find few of them or none, this can give them additional reason to believe the expert’s assertions. However, Hansen’s exploration makes clear that it is necessary to reflect upon the conditions that are to be included in a satisfactory argument scheme for good authority arguments. Steve Oswald and Thierry Herman present in ‘Give the standard treatment of fallacies a chance’ (Chap. 4) another take on the theoretical problem of how it can be explained that fallacies committed in argumentative discourse bypass or withstand critical testing than the one the pragma-dialecticians substantiated earlier in their Hidden Fallaciousness project (van Eemeren, Meuffels, & Garssen, 2012a, 2012b, 2015). In their research Oswald and Herman pay attention to the epistemic strength

1 A Variety of Contributions to Argumentation Theory

3

and ease of processing of fallacies and note that in argumentative discourse these cognitive parameters can be systematically manipulated to make fallacies rhetorically effective. In their expose they develop first a rationale for viewing rhetorical effectiveness in terms of foregrounding and backgrounding cognitive processes. Next they explain how these processes, and more in particular the evaluative ones, may by constrained decisively by the linguistic choices made in fallacious arguments. They illustrate their views in Chap. 4 through a discussion of a complex example of fallacious argumentation. In order to gain a better understanding of philosophical dialogues as a critical discussion, Federico López focuses in ‘Argument evaluation in philosophy’ (Chap. 5) on the evaluative strategies used by philosophers in their argumentative discourse, especially when they are accusing others of committing a fallacy. To illustrate the problems involved, López analyses the charge of committing a genetic fallacy. He provides a characterization of this fallacy and then argues that the accusation of committing a genetic fallacy should be understood pragma-dialectically as a case of (derailed) strategic manoeuvring (van Eemeren, 2010: 25–50, 93–128, 187–212). By revisiting the controversy over the genetic fallacy in the 1950s and 1960s, he presents some historical evidence in favour of this view. Next he gives an account of accusing the other party of committing a genetic fallacy as a mode of strategic manoeuvring and discusses some ways in which such strategic manoeuvring can derail. López concludes Chap. 5 with some general observations regarding the use of certain fallacies in philosophical controversies that are based on viewing the use of philosophical arguments in this way as cases of strategic manoeuvring. The special interest in philosophical argumentation shown by López is also manifest in several other contributions to this volume. They deal with such diverse topics as the “interrogative burden of proof”, “eudaimonistic argumentation”, and “community argument”. In the chapter devoted to the first of these topics the interrogative burden of proof in philosophical argumentation is related to the use of argumentative moves in dialogical sequences occurring in such discourse. In discussing the second topic of eudaimonistic argumentation in the next chapter it is argued that an account of argumentation contributing to human flourishing best serves the virtue epistemologists’ aim of realizing the intellectual good life. In dealing with community argument in the following chapter a communicative perspective is chosen in which certain worries about the possibilities of community argument are discussed that are motivated by empirical considerations. Chapter 6, ‘Dialogical sequences, argumentative moves and interrogative burden of proof in philosophical argumentation’, authored by Joaquin Galindo, consists of two parts. In the first part the concepts and terminology are introduced that Galindo uses in his argumentation. Galindo follows Rescher (1977) in distinguishing between “Initiating burden of proof” (I-burden) and “Evidential burden of proof” (E-burden). In his approach the obligation that exists in a dialectical situation to support an assertion with arguments has I-burden; counter-considerations and responses to counterargumentation have E-burden. In the second part of Chap. 6 Galindo introduces the notion of “philosophical argumentative move” in order to explain his views. He says

4

F. H. van Eemeren and B. Garssen

that philosophical argumentative moves can be found in chain arguments, counterarguments and even in argumentative plans and strategies. According to Galindo some of them are based on question-generating mechanisms that have the objective of allocating the burden of proof and generate various kinds of dialectical effects. His working hypothesis is that argumentative moves can be analyzed as dialogical sequences. In Chap. 6 he analyses three examples of philosophical argumentation (disclosing irrelevant distinctions, analytic dilemmas and pragmatic self-refutation) and concludes with the claim that the analysis can be easily extrapolated to other arguments. In ‘Eudaimonistic argumentation’ (Chap. 7) Andrew Aberdein explains that virtue theories of argumentation comprise several conceptually distinct projects. Among them, the pursuit of the fully satisfying argument—the argument that contributes to human flourishing—is in his view perhaps the boldest. The independently developed epistemic analogue of this project is “eudaimonistic virtue epistemology” (Brogaard, 2014). According to Aberdein, both projects stress the importance of widening the range of cognitive goals: in the first case beyond cogency and in the second case beyond knowledge. Because the right sort of community is indispensable for the cultivation of the intellectual virtues necessary to each of them, both projects emphasize social factors. In Chap. 7 Aberdein proposes to unify the two projects by arguing that the intellectual good life sought by eudaimonistic virtue epistemologists is best realized through the articulation of an account of argumentation that contributes to human flourishing. In ‘Worries about the prospects for community argument’ (Chap. 8) Dale Hample starts from the assumption that “community argument” takes place within a community of mutually-shared values and mutual respect. It may also invite outsiders to conjoin the arguers’ community during the argument, which produces a mutuality of concerns and respect. Hample thinks that this conjoining is not entirely natural, because our evolutionary heritage encourages us to valorize in-group members and to resist or attack outsiders. Our cognitive systems seem to be evolutionary equipped with the possibility to make certain information accessible once we are engaged in arguing while other information is made structurally less accessible. According to Hample, these evolutionary endowments interfere with treating an outsider as an intimate and reasoning properly on unfamiliar matters. Other worries mentioned by Hample relate to fundamental human impulses described in social psychology that seem to involve ethical deficits, at least when viewed from the perspective of a discipline committed to full and free argumentation as the route to personal and social improvement. The lines of persuasion utilized in the processes concerned are designed to close off argumentation and make social reasoning less likely to take place and less productive. The persuasive techniques then at work are aimed at suppressing counter-argumentation by discouraging people from thinking deeply about persuasive messages. According to Hample, the best known of these techniques is “nudging” (Johnson et al., 2012), which has become a public policy adopted by a great many national governments and is criticized for its ethics. Other persuasive techniques used for the same purpose mentioned by Hample are distraction, multitasking, narrative persuasion, using visual images, and situating

1 A Variety of Contributions to Argumentation Theory

5

persuasion in computer games and virtual reality. In Hample’s view the argumentation community needs to develop pedagogies to counteract these techniques. Next, the volume contains two contributions devoted to the role of counterarguments in evaluating the quality of argumentation. One of these chapters focuses on testing the strength of warrants used in argumentation by assessing the extent to which they resist rebutting by counter-arguments that are pertinent to the institutional context in which the argumentation is advanced. The other chapter concentrates on distinguishing between different types of counter-arguments and explaining the consequences of the distinctions that are made for the analysis of the structure of argumentation. In ‘Assessing connection adequacy for arguments with institutional warrants’ (Chap. 9) James Freeman develops an account of the use of Toulminian warrants backed by laws. He explains how the conclusions he has reached in this endeavour can be applied to institutional warrants in general. According to Freeman, an institutional warrant is always backed by rules (Freeman, 2005). An appraisal of the strength of the warrant goes in his view from the rules to the warrant and involves judging its strength in resisting rebutting values of relevant variables. Whether the warrant has sufficient strength for cogency depends on the requirements of the institution that is backing the warrant—which may lead to differences such as those between preponderance of evidence and proof beyond a reasonable doubt. Freeman’s conclusion is that understanding the strength of a warrant always involves characterizing rebuttal avoidance. A warrant is sufficient for cogency when it avoids rebutting or when the argument in which it is used involves counter-rebuttals at each level below the threshold for rebuttal indifference. According to Hubert Marraud in ‘On the logical ways to counter an argument’ (Chap. 10), the concept of counter-argument lies at the heart of dialectic as the study of “oppositions between arguments” (Marraud, 2015). In Chap. 10 he is out to give a suitable definition of counter-argument, to establish a typology of counter-arguments and to explore the consequences of his views for the theorizing about argumentation structure. Marraud confines himself in this enterprise to counter-arguments in which an argument is criticized by other arguments on account of its logical properties. Because in his definition a counter-argument is a reason to reject another argument, there is a clear difference between a counter-argument presenting a reason against the cogency of an argument and a con argument or a reason against a claim. In a counter-argument one can argue that the argument that is advanced is not cogent, that certain premises of the argument are not true or not acceptable, that the premises of the argument are not adequately connected to the conclusion or that ultimately the conclusion is false or unacceptable. Marraud distinguishes accordingly between four main kinds of counter-arguments: “dismissal”, “objection”, “rebuttal” and “refutation”. An argumentation (or argument) structure consists of a combination of arguments with an explicit argumentative orientation—positive, negative or neutral. Advancing counter-arguments always involves conjoining the criticized argument and the critical argument into one and the same argumentation. Marraud therefore concludes that the standard catalogue of argumentation structures, which includes (combinations

6

F. H. van Eemeren and B. Garssen

of) multiple, coordinative and subordinative argumentation, does not suffice in representing the argumentation structure of argumentation involving counter-arguments in diagrams, but needs to be supplemented with counter-argument structures covering objection, rebuttal and refutation. Thus the proposed classification of counterarguments is in his opinion also a classification of argumentation structures. There are also three chapters of the volume that are devoted to the argumentative use of language and the problems involved in its reconstruction. The first one makes a case for regarding certain questions argued for in argumentative practice as genuine questions and indicates how the inferences made in the arguments concerned can be evaluated. The second one zooms in on the role of the Dutch adverb “straks” [“next”] in the argumentative use of a type of utterances in which arguers express their negative emotional stance. The third one explains how the abductive reasoning can be reconstructed and assessed in a reasoned way that is involved in conveying in argumentative discourse an intent that deviates from the intent that seems to be expressed. As can be verified by a Web search using phrases containing conclusion indicators and interrogative particles such as “so why” or “how then” as search terms, in argumentative practice people do sometimes appear to argue for questions. In ‘Arguing for questions’ (Chap. 11) David Hitchcock claims that the texts concerned are really arguments and their conclusions are genuine questions. He contends that people typically argue for questions to stimulate interest in discovering the answer to the question concerned or to challenge the addressees or a third party to explain their behaviour. According to Hitchcock, the “inferential erotetic logic” of the Polish logician Wi´sniewski (1996, 2013) provides a basis for evaluating the inferences from statements to questions that are made in such arguments. For an inference from one or more statements to a question to be valid, the statements and the context in which they are advanced must entail that the question has a true answer without entailing that some particular answer is true. In addition, there is the pragmatic rather than semantic constraint that there must be a point to asking the question, such as the addressees’ ignorance of the correct answer(s) to this question. In ‘Expressives in argumentation: The case of apprehensive straks (“shortly”) in Dutch’ (Chap. 12) Ronny Boogaart concentrates on the argumentative use of a certain “expressive construction” conveying a negative emotion that typically seems to involve a derailed strategic manoeuvre. As a case in point he discusses some pertinent uses of the Dutch adverb straks in this expressive construction. In some of them pragmatic argumentation is advanced against future consequences that the arguer values negatively; in some others analogy argumentation is used in which a standpoint is attacked by comparing it with an allegedly related extreme standpoint that is clearly untrue or absurd. The question to be answered is, according to Boogaart, what the rhetorical advantage is of strategic manoeuvring by making argumentative moves that are so intimately related to the fallacies of incorrect causal argumentation and false analogy. In his view, presenting an argument by means of the expressive construction he discusses functions as an “immunization” technique (van Eemeren & Grootendorst, 1992: 119). First, by giving this expressive shape to the argumentation it is suggested that there are feelings involved which are beyond the arguer’s control

1 A Variety of Contributions to Argumentation Theory

7

so that the burden of proof is evaded. Second, by being unreasonable in such an overt way it is suggested that there is no derailment because the relevant discussion rule is flouted rather than really violated. According to Antonio Duarte in ‘Argumentative abduction in the interpretation process’ (Chap. 13) abductive reasoning necessarily plays an important role in the interpretation of deviations of intent from the intent expressed by the “standard” or ordinary use of language. To achieve the desired effect in the audience, the speaker uses in such cases usually some “mysterious” words and relies on the hearer’s abductive ability to solve the enigma. As Duarte explains, this abductive ability is mostly based on the interlocutors’ “knowledge of the common dialogic framework”. Both speaker and listener (or writer and reader) have to apply this kind of abductive reasoning in order to make their communication successful. When the exchange they are having is understood as an argumentative process, the kind of “guessing” involved can be understood and evaluated in a reasoned way. As a practical case in point, Duarte analyses in Chap. 13 an ironic utterance from a pragma-dialectical perspective. In doing so he proposes to evaluate the kinds of hypotheses developed in the interpretation process in an argumentative way as responses to certain critical questions ensuing from the theoretical norms described in extended pragma-dialectics. There is also a section of the volume consisting of three chapters dealing with conductive arguments. In these contributions this topic is treated in different ways. In the first contribution it is argued that calling an argument a “conductive argument” can be a mistake. In the second contribution it is claimed that the two most prominent approaches to conductive argument are both lacking and it is explained why. The third contribution claims to offer a solution to the problems that, in the author’s view, the existing approaches he discusses cannot cope with. Isabela Fairclough tackles in Chap. 14 the question ‘Is “conductive argument” a single argument?’. Focusing on a particular kind of conductive argument, i.e. a “pro/con” argument intended to support a practical conclusion, she argues that “conductive argument” is a category mistake: if the term is meant to designate a single argument, with one conclusion, there is in her view no such thing as a conductive argument. As a structure, what appears to be a conductive argument is in fact one of the two possible outcomes of the deliberative activity of critically testing alternative proposals for action. More precisely, it is a recapitulation or summary of a process of critical questioning in which both counter-considerations and reasons in favour of a practical conclusion have been advanced but no decisive objections against it have emerged, so that the conclusion has withstood criticism. Here an important distinction is being drawn between two kinds of objections that can be raised against a proposal: counter-considerations which do not refute it and decisive objections which do. Saying that there are no decisive objections against a proposal means that is does not follow conclusively that the proposal should be abandoned. According to Fairclough, the positive conclusion together with all the reasons cited in its favour and all the reasons cited against it (the counter-considerations) will be virtually indistinguishable from a so-called “conductive argument”. If in the argumentative process the positive conclusion does not survive the criticism that is

8

F. H. van Eemeren and B. Garssen

advanced, the potential conductive argument disintegrates, collapsing into a deductive argument in favour of the negative conclusion. In ‘On the logical reconstruction of conductive arguments’ (Chap. 15) Yun Xie discusses two important approaches to the logical reconstruction of conductive arguments. The first one is the “supplementation-of-on-balance-premise approach”, which is advocated by informal logicians such as Govier (2011), Hansen (2011) and Blair (2017). The second one is the “warrant-reformulation approach”, which is proposed by Freeman (2011) and Bermejo-Luque (2017). According to Xie, in the versions of these two approaches that have been developed so far neither of them is satisfactory. On the one hand, reconstructing conductive arguments by adding an on-balance premise as happens in the first approach is problematic, because such a reconstruction is based on a presumption that is still in need of further justification and the interactions between the reasons for and against are in this approach simplified too much. On the other hand, integrating counter-considerations into the warrant of a conductive argument, as would happen in the other approach, leads to a reconstruction which is indifferent to the linguistic features of the arguer’s conduct and thus fails to give full credit to the arguer’s use of language. Yanlin Liao observes in ‘The legitimacy of conductive arguments’ (Chap. 16) that recent disputes about whether conductive arguments are a distinct type of argument have given rise to a “legitimacy crisis of conduction”. Adler (2013) once argued that the definition of conduction implies two incompatible claims which make conduction impossible. The crux of his position can be summarized in the question of how it can be possible that the conclusions of conductive arguments are unqualified while the negative considerations that are involved remain viable. Xie and Xiong (2013) and Xie (2017) have provided a rhetorical solution to the problem by claiming that the negative considerations have only a rhetorical function so that in the arguments they are not at all viable. Blair (2016), however, contends that the negative considerations have a logical function and can weaken the strength of the argument. In Chap. 16 Liao reconstructs first Jonathan Adler’s argument and the variant of it that he has distinguished. Next he presents his criticism of the rhetorical solution of Yun Xie and Minghui Xiong as well as the logical solution of Anthony Blair. Although Blair’s logical solution can resolve the problem raised by Adler, it does not offer a complete defence of the legitimacy of conduction. According to Liao, the logical function of the negative considerations needs to be fully justified in order to defend the legitimacy of conduction in a better way. In making an effort to do so he provides a new perspective of argument evaluation in the case of conduction in which he argues that the presence of the negative considerations can change the original criteria of argument evaluation into new criteria. In Liao’s view, the logical function of counter-considerations therefore lies in changing the criteria of argument evaluation. He concludes that in this way the legitimacy of conduction can be fully justified. From argument schemes to argumentative relations in the wild concludes with a contribution pertaining to an area of research that is these days vital to the applicability of argumentation theory: the automatic detection of relevant elements and aspects of argumentative discourse. This research into “argument mining” (Lippi & Torroni,

1 A Variety of Contributions to Argumentation Theory

9

2016) focuses primarily on the detection of arguments, components of arguments and relations between arguments in automated corpora using a combination of natural language processing and machine learning techniques. The chapter included in this volume concentrates on so-called Relation-based Argument Mining. In ‘Deploying machine learning classifiers for argumentative relations “in the wild”’ (Chap. 17) Oana Cocarascu and Francesca Toni describe two experiments concerning the deployment of various Relation-based Argument Mining classifiers in order to extract argumentative relations of attack and support from a dialogue and a short text. They claim that outputs of the classifiers can be used to construct directed graphs representing the relations between arguments from any given text. These graphs can then be deployed to support a number of applications, varying from capturing and analysing debates in social media to supporting decision making. Cocarascu and Toni experiment with various machine learning techniques (Support Vector Machines, Random Forests, Long-Short Term Memory networks) to determine argumentative relations. They trained the classifiers in dealing with a data set consisting of pairs of texts annotated as “attack”, “support” or “neither attack nor support”. By comparing the argumentative relations extracted by making use of trained classifiers with the ones extracted by human annotators, they are able to show that there is no overall winner as to which machine learning classifier is better suited for this task: Long-Short Term Memory networks performed best for the first experiment and Random Forests for the second experiment. This suggests that different techniques might be better suited for different types of texts. The general conclusion Cocarascu and Toni draw from their analysis is that the search for an all-encompassing method is still open.

References Adler, J. E. (2013). Are conductive arguments possible? Argumentation, 27(2), 245–257. Bermejo-Luque, L. (2017). The appraisal of conductions. In S. Oswald & D. Maillat (Eds.), Argumentation and inference (pp. 1–18). London: College Publications. Blair, J. A. (2016). A defense of conduction: A reply to Adler. Argumentation, 30(2), 109–128. Blair, J. A. (2017). In defence of conduction: Two neglected features of argumentation. In S. Oswald & D. Maillat (Eds.), Argumentation and inference (pp. 29–44). London: College Publications. Brogaard, B. (2014). Towards a eudaimonistic virtue epistemology. In A. Fairweather (Ed.), Virtue epistemology naturalized. Bridges between virtue epistemology and philosophy of science (pp. 83–102). Cham, Switzerland: Springer. Freeman, J. B. (2005). Acceptable premises. An epistemic approach to an informal logic problem. Cambridge: Cambridge University Press. Freeman, J. B. (2011). Evaluating conductive arguments in light of the Toulmin model. In J. A. Blair & R. H. Johnson (Eds.), Conductive arguments. An overlooked type of defeasible reasoning (pp. 127–144). London: College Publications. Goldberg, S. C. (2011). If that were true I would have heard about it by now. In S. C. Goldman & D. Whitcomb (Eds.), Social epistemology: Essential readings (pp. 92–108). New York: Oxford University Press.

10

F. H. van Eemeren and B. Garssen

Govier, T. (2011). Conductive arguments: Overview of the symposium. In J. A. Blair & R. H. Johnson (Eds.), Conductive arguments. An overlooked type of defeasible reasoning (pp. 262–276). London: College Publications. Hansen, H. V. (2011). Notes on balance-of-consideration arguments. In J. A. Blair & R. H. Johnson (Eds.), Conductive arguments. An overlooked type of defeasible reasoning (pp. 31–51). London: College Publications. Johnson, E. J., Shu, S. B., Dellaert, B. G. C., Fox, C., Goldstein, D. G., Häubl, G., et al. (2012). Beyond nudges: Tools of a choice architecture. Mark Lett, 23, 487–504. Lippi, M., & Torroni, P. (2016). Argumentation mining. State of the art and emerging trends. ACM Transactions on Internet Technology, 16(2), 10. Marraud, H. (2015). Do arguers dream of logical standards? Arguers’ dialectic vs. arguments’ dialectic. Revista Iberoamericana de Argumentación, 10. Rescher, N. (1977). Dialectics. A controversy-oriented approach to the theory of knowledge. Albany: State University of New York Press. van Eemeren, F. H. (2010). Strategic maneuvering in argumentative discourse. Extending the pragma-dialectical theory of argumentation. Amsterdam/Philadelphia: John Benjamins. Argumentation in Context 2. van Eemeren, F. H., Garssen, B., & Meuffels, B. (2012a). Effectiveness through reasonableness. Preliminary steps to pragma-dialectical effectiveness research. Argumentation, 26(1), 33–53. van Eemeren, F. H., Garssen, B., & Meuffels, B. (2012b). The disguised abusive ad hominem empirically investigated. Strategic maneuvering with direct personal attacks. Thinking & Reasoning, 18(3), 344–364. Special issue Reasoning and Argumentation. van Eemeren, F. H., Garssen, B., & Meuffels, B. (2015). The disguised ad baculum fallacy empirically investigated. Strategic maneuvering with threats. In F. H. van Eemeren & B. J. Garssen (Eds.), Scrutinizing argumentation in practice (pp. 313–326). Amsterdam/Philadelphia: John Benjamins. Argumentation in Context 9. van Eemeren, F. H., & Grootendorst, R. (1992). Argumentation, communication, and fallacies. A pragma-dialectical perspective. Hillsdale, NJ: Lawrence Erlbaum. Wi´sniewski, A. (1996). The logic of questions as a theory of erotetic arguments. Synthese, 109(1), 1–25. Wi´sniewski, A. (2013). Questions, inferences and scenarios. London: College Publications. Xie, Y., & Xiong, M. (2013). Commentary on J. A. Blair’s ‘Are conductive arguments really not possible?’. In D. Mohammed & M. Lewi`nski (Eds.), Virtues of argumentation. Proceedings of the 10th international conference of the Ontario Society for the Study of Argumentation (pp. 1–6). Windsor, ON: OSSA. Xie, Y. (2017). Conductive argument as a mode of strategic maneuvering. Informal Logic, 37, 2–22.

Chapter 2

Argument Schemes: Extending the Pragma-Dialectical Approach Frans H. van Eemeren and Bart Garssen

2.1 The Notion of Argument Scheme In the last two decades various theoretically-oriented publications have appeared 1 about argument schemes. Not much exposure however has been given lately to the pragma-dialectical perspective on argument schemes and the way in which it has 2 developed since the late 1970s. Recently the two of us have started a project aimed at explaining the pragma-dialectical theory of argument schemes and extending it with new insights. This paper is intended to be the first instalment of a more encompassing series of studies that is to result in a monograph in which a complete overview will be given of the current state of affairs in the pragma-dialectical treatment of argument 1 See

for instance Walton, Reed and Macagno (2008), Lumer (2011) and Wagemans (2016). idea of argument schemes and critical questions in a dialectical testing procedure was introduced in van Eemeren, Grootendorst and Kruiger (1978). Van Eemeren and Kruiger (1987) give an account of the identification of argument schemes. In van Eemeren and Grootendorst (1992) the pragma-dialectical typology of argument schemes is elaborated, while van Eemeren and Grootendorst (2004) give a detailed account of the pragma-dialectical testing procedure. Garssen (1997) presents an exploration of possible subtypes and variants based on a comparison of the pragmadialectical typology and other prominent typologies of argument types. Van Eemeren and Garssen (2014) explain the pragma-dialectical view of analogy argumentation and its role in prototypical argumentative patterns in political argumentation. Van Eemeren (2018: 45–49) provides an update of the pragma-dialectical treatment of argument schemes.

2 The

F. H. van Eemeren (B) · B. Garssen Speech Communication, Argumentation Theory and Rhetoric, ILIAS, University of Amsterdam, Amsterdam, The Netherlands e-mail: [email protected] B. Garssen e-mail: [email protected] F. H. van Eemeren Leiden University, Leiden, The Netherlands © Springer Nature Switzerland AG 2020 F. H. van Eemeren and B. Garssen (eds.), From Argument Schemes to Argumentative Relations in the Wild, Argumentation Library 35, https://doi.org/10.1007/978-3-030-28367-4_2

11

12

F. H. van Eemeren and B. Garssen

schemes and their categorization in a theoretically motivated and empirically justified typology. In the pragma-dialectical perspective, argumentation is aimed at resolving a difference of opinion about an evaluative, prescriptive or descriptive standpoint. Based on the starting points accepted as their point of departure by the parties in the difference, the standpoint at issue is in argumentation defended by advancing one or more reasons in its support. The reasons that are advanced in argumentation are intended to offer an informal justification of the acceptability of the standpoint at issue, not a definitive proof of its truth.3 When a standpoint can be proven true by an immediate empirical check or a demonstration that it follows logically from true premises, doing so will suffice and there is no need for argumentation—or at most this proof could be presented as an irrefutable argumentation. When the truth of a standpoint can be shown beyond any doubt by presenting a modus ponens-like formal derivation of the standpoint, the only step that needs to be taken in evaluating the argumentation thus advanced is checking the logical validity of the reasoning involved. However, in ordinary argumentation the reasoning is as a rule not explicitly presented in this way, so that carrying out such a check will usually not be possible – or can only be accomplished in an extremely artificial way. In ordinary argumentation the acceptability of a standpoint is in principle defended by linking the propositional content of the argumentation by means of a particular justificatory principle to the standpoint at issue. This means that the acceptability of the standpoint at issue depends on the suitability and correctness of the use of the argument scheme brought to bear in applying this justificatory principle. There are various types of argumentation that can be used in defending the acceptability of a standpoint, each of which is characterized by having a particular argument scheme. Each argument scheme represents a particular justificatory relationship between a reason (or cluster of interdependent reasons) and a standpoint that is supposed to legitimize the transfer of acceptability from the reason (or cluster of interdependent reasons) advanced to the standpoint that is defended. The various argument schemes that can be used and the way in which their use is to be evaluated are in the pragma-dialectical view part of the joint starting points that are in principle by intersubjective agreement explicitly or implicitly established at the opening stage of a critical discussion aimed at resolving a difference of opinion on the merits. When the argumentation advanced in defence of a standpoint consists of a plurality of reasons that are in some combination or other advanced in support of a standpoint, it has a complex argumentation structure. Since each individual justification of a standpoint has its own argument scheme, whether it consists of a single or a coordinative argumentation (as in the case of the use of interdependent reasons), such complex argumentation may involve the use of more than one argument scheme. This means that, in principle, the argument schemes that are used in complex argumentation do not automatically pertain to the argumentation as a whole, but to its various justificatory constituents. Although in 1978, when we started to make 3 Instead

of an informal justification, an informal refutation can also be offered. For the sake of brevity, we will refrain from adding this all the time.

2 Argument Schemes: Extending the Pragma-Dialectical Approach

13

use of this concept (van Eemeren et al., 1978), we had initially opted for using the term argumentation scheme, we therefore later decided for the sake of clarity to give preference to the term argument scheme (e.g. van Eemeren & Grootendorst, 2004).

2.2 Intersubjective Procedures for Evaluating Argumentation According to the “Munchhausen trilemma” sketched by critical rationalist Hans Albert, there are three ways in which providing a justification of a standpoint will finally always come to a dead end. Two of them, circularity and an infinite regress, are indeed fatal. However, the third option that Albert distinguishes, breaking off the justification at an arbitrary point, is in our view not inescapable. If the justification process is ended when a starting point has been reached that is recognized by both parties, the justification is not concluded in an arbitrary way, but has a pragmatic basis in well-considered intersubjective agreement. This reliance on existing agreement, which may have been established explicitly or correctly presumed, is in fact quintessential to any serious conduct of argumentation. It is the very reason why in the pragma-dialectical approach to argumentative discourse the “opening stage”, where the procedural and material starting points of the resolution process are determined, is considered vital to resolving a difference of opinion on the merits. The pragma-dialectical rules for resolving a difference of opinion on the merits include a set of procedures for evaluating argumentation that are supposed to be intersubjectively agreed upon in the opening stage of a critical discussion (van Eemeren & Grootendorst, 2004: 135–157). The identification procedure involves determining whether a proposition called into question in resolving the difference of opinion is identical to any of the propositions which may be regarded jointly accepted starting points. If a proposition may be regarded to be part of the point of departure that has been accepted at the opening stage of the discussion, it may not be called into question in the argumentative exchange of the ongoing discussion. In order to allow for new information to be used in the argumentative exchange that is not already included in the starting points, the parties may in the opening stage agree to leave room for sub-discussions in which it is determined whether a proposition that was initially not agreed upon can be accepted as a starting point in the second instance. Next there is the inference procedure, which is aimed at determining whether in cases in which the reasoning is fully externalized the reasoning “proposition involved in the argumentation, therefore proposition involved in the standpoint” presented by the protagonist is logically valid as it stands. If the reasoning is not completely externalized, so that the argumentation cannot be logically valid as it stands, as is in argumentative practice generally the case, the question is whether the argument schemes that are brought to bear in the argumentation are admissible to both parties and have been used correctly in the case concerned. If it first needs to be determined which argument scheme has been employed before

14

F. H. van Eemeren and B. Garssen

this can be decided, then the explicitization procedure needs to be followed, which is for this purpose added to the available pragma-dialectical tools. To check whether a particular argument scheme has been used correctly, the testing procedure must be carried out. This procedure consists of asking the critical questions appropriate for checking the correctness of the use of a particular argument scheme. Each argument scheme gives cause to different critical questions, which open up different kinds of dialectical routes. For a conclusive defence of the standpoint, both the propositional content of the argumentation that is advanced and its justifying force must have been defended successfully in accordance with the relevant evaluation procedures. For a conclusive attack on the standpoint, either the propositional content of the argumentation or its justifying force must have been attacked successfully in accordance with the relevant evaluation procedures. In the present context it is worth repeating that the intersubjective agreements that are part of the joint starting points established in the opening stage of a critical discussion aimed at resolving a difference of opinion on the merits are supposed to include the various argument schemes that can be used and the way in which their use is to be evaluated.

2.3 The Pragma-Dialectical Typology of Argument Schemes In order to be able to carry out the pragma-dialectical testing procedure, a problemvalid inventory of argument schemes is required. This means that the inventory cannot be just a taxonomy but must be a theoretically-motivated typology that involves a categorisation relating to the properties of argumentation that are relevant to its evaluation. It stands to reason that in the pragma-dialectical theory of argumentation the rationale for distinguishing between the various categories of argument schemes in a general classification has a pragmatic as well as a dialectical dimension.4 The pragmatic dimension relates to the kind of justificatory principle that legitimizes in an argument scheme the transfer of acceptance from the reason advanced to the standpoint that is defended. This is in the pragma-dialectical view not a formal principle, as it is in establishing logical validity, but a pragmatic one, based on human experience, i.e. grounded in the practical justificatory experiences of arguers in ordinary argumentative discourse.5 The dialectical dimension relates to the dialogical evaluation procedure associated with the argument scheme that is used, i.e. to the critical questions that are to be answered satisfactorily in order to legitimize the use of the argument scheme concerned. When taken together, these two dimensions 4 See

Garssen (2001) for an overview of other kinds of classifications of argument schemes. is probably the same grounding in “native analytic categories” as Doury (2018) speaks of. This pragmatic basis, which is similar to that of the “ortho-language” of the logical propaedeutic of Kamlah and Lorenzen (1984), manifests itself in the various expressions by which the argument schemes are indicated in ordinary language (van Eemeren & Kruiger, 1987).

5 This

2 Argument Schemes: Extending the Pragma-Dialectical Approach

15

constitute the principium divisionis underlying the typology of argument schemes that is in pragma-dialectics presumed to be part of the intersubjectively accepted starting points for a critical discussion. The three main categories of argument schemes distinguished in pragma-dialectics are “symptomatic” argumentation (also known as “sign” argumentation), “comparison” argumentation (also known as “resemblance” argumentation) and “causal” argumentation (also known as “consequence” argumentation) (van Eemeren et al., 1983: 137–141; van Eemeren & Grootendorst, 1992: 94–102). Symptomatic argumentation, to start with, is a type of argumentation in which an argument scheme is used that is based on the pragmatic principle of something being symptomatic of something else, i.e. the one being a token or a sign of the other. Symptomatic argumentation involves a relation of concomitance between the reason advanced and the standpoint defended (e.g. “Pinchao is a Chinese [and it is goes with Chinese people that they are diligent], so he is bound to be diligent”). Comparison argumentation is a type of argumentation in which an argument scheme is used that is based on the pragmatic principle of something being comparable to something else, i.e. the one resembling or being similar to the other. Comparison argumentation involves a relation of comparability between the reason advanced and the standpoint defended (e.g. “Camera surveillance in the Amsterdam metro will be effective because it is also effective in the London underground [and the situation in Amsterdam is comparable to the situation in London]”). Causal argumentation is a type of argumentation in which an argument scheme is used that is based on the pragmatic principle of something being causal to or consequential of something else, i.e. the one being instrumental to or leading to the other. Causal argumentation involves a relation of instrumentality or consequentiality between the reason advanced and the standpoint defended (e.g. “Because Alfonso has exercised very long [and exercising very long leads to tiredness], he must be tired”). Because each of the argument schemes calls out its own set of critical questions, the three categories of argument schemes thus distinguished are associated with specific dialectical routes in resolving a difference of opinion on the merits. The differences between the dialectical routes instigated by the use of symptomatic argumentation, comparison argumentation and causal argumentation are in the first place determined by the basic critical question connected with the category of argument schemes concerned, which relates to the (usually unexpressed) bridging premise. The basic critical question associated with symptomatic argumentation is whether what is claimed in the standpoint (Y) is indeed a sign of what is stated in de reason advanced (X) (or whether what is stated in the reason (X) is indeed a token of what is claimed in the standpoint (Y)). In argumentation of this type, protagonist P defends standpoint Y (e.g. Chinese Pinchao is diligent [PD]) against antagonist A’s doubt Y? (e.g. [PD?]) by advancing symptomatic argumentation X (e.g. Chinese people are diligent [CD]) and A responds critically by asking the basic critical question connected with symptomatic argumentation (e.g. whether being diligent is indeed characteristic of Chinese people [C//D?]), which will lead to an answer by P (e.g. [D//C: OK]) and may be followed by further discussion.

16

F. H. van Eemeren and B. Garssen

A simplified dialectical profile of symptomatic argumentationthat only includes the just indicated dialectical route instigated by the basic critical question looks as follows: 1. 2. 3. 4. 5.

P: Standpoint: Y [PD] A: Y? [PD?] P: Symptomatic argumentation: X [DC] A: Basic critical question: Is Y symptomatic of X (Y//X)? [D//C?] P: Answer to basic critical question: Y//X: OK [D//C: OK] (which may be followed by further discussion).

The basic critical question associated with the use of comparison argumentation is whether what is claimed in the standpoint (Y) is indeed comparable to what is stated in the reason advanced (X) [or whether what is stated in the reason (X) is indeed similar to what is claimed in the standpoint (Y)]. In argumentation of this type protagonist P defends standpoint Y (e.g. Late-comer Vahid should not be allowed to participate [~VP < VL]) against antagonist A’s doubt Y? (e.g. [(~VP < VL)?]) by advancing comparison argumentation X (e.g. Other people who did not meet the deadline in the past were not allowed to take part [~OP < OL]), to which A responds by asking the basic critical question connected with comparison argumentation Y = X? (e.g. whether your being late is indeed comparable to other people not meeting the deadline in the past [VL = OL?), which leads to an answer: Y = X: OK (e.g. [VL = OL: OK]) and may be followed by further discussion. A simplified dialectical profile of comparison argumentation including only this dialectical route instigated by the basic critical question is as follows: 1. 2. 3. 4. 5.

P: Standpoint: Y [~VP < VL] A: Y? [(~VP < VL)?] P: Comparison argumentation: X [~OP < OL] A: Basic critical question: is Y comparable with X: (Y = X)? [VL = OL?] P: Answer to basic critical question: Y = X: OK [VL = OL: OK] (which may be followed by further discussion).

The basic critical question associated with causal argumentation is whether what is stated in the reason that is advanced (X) leads to what is claimed in the standpoint (Y) [or whether what is claimed in the standpoint (Y) indeed results from what is stated in the reason that is advanced (X)]. In argumentation of this type protagonist P defends standpoint Y (e.g. Alfonso must be tired [AT]) against antagonist A’s doubt Y? (e.g. [AT?]) by advancing causal argumentation X (e.g. Alfonso has exercised very long [AE]), to which A responds by asking the basic critical question of causal argumentation (e.g. whether exercising very long does indeed always lead to great tiredness [(T < E)?] or, more precisely, [((x)xT < (x)xE)?], which leads to an answer (e.g. [T < E: OK] or [(x)xT < (x)xE: OK]) and may be followed by further discussion. A simplified dialectical profile of causal argumentation including only this dialectical route instigated by the basic critical question is as follows:

2 Argument Schemes: Extending the Pragma-Dialectical Approach

17

1. 2. 3. 4.

P: Standpoint: Y [AT] A: Y? [AT?] P: Causal argumentation: X [AE] A: Basic critical question: Does X lead to Y? (Y < X?) [T < E? or ((x)xT < (x)xE)?] 5. P: Answer to basic critical question: Y < X: OK [T < E: OK or (x)xT < (x)xE: OK] (which may be followed by further discussion).

2.4 Critical Questions: Pragmatic Argumentation as a Case in Point Next to the basic critical question connected with the category of argument schemes that is brought to bear in the argumentation, in the testing procedure carried out in evaluating the argumentation there are always still other questions that the antagonist may be supposed to ask. Which other critical questions are relevant depends in the first place on the type of argumentation involved. Let us, in explaining what kind of further critical questions can be relevant in evaluating argumentation, by way of example concentrate on “pragmatic” argumentation, a prominent subtype of causal argumentation. The idea underlying pragmatic argumentation is that we must do something because it leads to something we want to happen. Put more precisely, in pragmatic argumentation the prescriptive standpoint that a certain action should be carried out is defended by pointing out that carrying out this action leads to a certain desirable result—or, in the negative version of pragmatic argumentation, that a certain action should not be carried out by pointing out that carrying out this action leads to a certain undesirable result. The positive version of the argument scheme brought to bear in pragmatic argumentation can be specified as follows: 1. Action X should be carried out 1.1 Action X leads to desirable result Y (1.1’) (If action X leads to a desirable result such as Y, X must be carried out). All critical questions asked in carrying out the testing procedure in order to evaluate the use of argument schemes pertain to the argumentation as it has been externalized by means of the intersubjective explicitization procedure. The basic critical question asked in this procedure always concerns the relationship established by the use of the category of argument schemes concerned between the reason that is advanced and the standpoint that is defended. When, as in the case of pragmatic argumentation, a causal argument scheme is used, the basic critical question therefore is: (a) Does action X indeed lead to result Y? As the bridging premise externalized in the explicitization of the argumentation makes clear, when the subtype of pragmatic

18

F. H. van Eemeren and B. Garssen

argumentation is used the next relevant question relating to this basic critical question will be: (b) Must actions that lead to a desirable result Y always be carried out?6 Other relevant critical questions included in the testing procedure pertain to the non-bridging premise of the argumentation as it has been externalized in the explicitization procedure (or to the non-bridging premises in cases where the argumentation is coordinative). Some of these critical questions concern the acceptability of such a premise or vital presuppositions involved in this premise. In the case of pragmatic argumentation the critical question about the acceptability of the explicit non-bridging premise (Does action X indeed lead to result Y?) has in fact already been asked because it happens to be identical with the basic critical question for causal argumentation in general. However, in the case of pragmatic argumentation another relevant critical question relating to the non-bridging premise concerns a crucial presupposition involved in this premise: (c) Is result Y indeed desirable? There are also some further critical questions that are relevant to testing the acceptability of the use of pragmatic argumentation which are in a more indirect way connected with the critical questions just distinguished. They relate to the specific point of pragmatic argumentation that carrying out an action is justified by its desirability and refer to other possibilities or options that need to be taken into account when deciding about the adequacy of the argumentation. One of these critical questions pertains to the explicit premise “Action X leads to desirable result Y” (1.1): (d) Would another result not be even more desirable than Y? Two other relevant critical questions pertain in different ways to the unexpressed bridging premise “If action X leads to a desirable result such as Y, X must be carried out” (1.1’): (e) Does action X not have unavoidable undesirable side-effects?; (f) Could result Y not be achieved more easily or more economically by other actions? If in the argumentative discourse any of the critical questions (a)-(f) is anticipated or answered, the argumentation involved becomes automatically complex. Then the difference of opinion at issue can no longer be said to have been resolved by argumentation that is “pragmatic” in the specific sense that it is conclusive by putting the difference in one go to an end. If, for instance, in view of critical question (c) relating to pragmatic argumentation, the desirability of the result that will be reached needs to be motivated since this desirability is not beyond doubt, the argumentation that has been advanced loses its pragmatic force of instantaneous effectiveness. In such a case the argumentation remains, of course, causal but turns from straightforward pragmatic argumentation into “complex pragmatic (problem-solving) argumentation”, in which the initial pragmatic argumentation is embedded in more complex argumentation. As always in argumentative discourse, it then depends on the argument schemes that are brought to bear in the argumentation advanced in answering 6 Unlike

in other subtypes of causal argumentation, the bridging premise relates in the case of pragmatic argumentation, due to the complex nature of this subtype, only indirectly to the basic critical question of causal argumentation: the causal relation at issue in the basic critical question is in pragmatic argumentation presumed. In evaluating the use of pragmatic argumentation, just as in evaluating the use of other subtypes of causal argumentation, the basic critical question is to be answered first before it makes sense to turn to the critical questions specifically relating to this particular subtype.

2 Argument Schemes: Extending the Pragma-Dialectical Approach

19

the critical questions associated with pragmatic argumentation which further critical questions need to be answered in the continuation of the discourse.

2.5 Subtypes and Variants of Argumentation As we have indicated, pragmatic argumentation is a subtype of the general category of causal argumentation in which the basic critical question applying to causal argumentation is complemented by an additional critical question that focuses on the presupposition crucial to the justificatory point of this subtype that the desirability of the result of the action justifies carrying it out. The basic critical question of causal argumentation (“Does action X indeed lead to result Y?”) is in the case of pragmatic argumentation supplemented by the additional critical question “Must actions that lead to a desirable result Y always be carried out?”. Such subtypes cannot only be distinguished within the category of causal argumentation but also within the categories of comparison argumentation and symptomatic argumentation (van Eemeren & Grootendorst, 1992: 97). Like in the case of the main categories, the rationale for distinguishing between the various subtypes is both pragmatic and dialectical. This means that each subtype should be characterized by relying on a specific justificatory principle based in human experience and leading to a different, i.e. uniquely specified, set of critical questions in the intersubjective testing procedure. By avoiding any differentiations in which these two preconditions have not been fulfilled our typology lives up to the old adage that “a difference that makes no difference is no difference”. Among the other subtypes of causal argumentation are—to name just a few— argumentation from cause to effect, argumentation from effect to cause and argumentation from means to goal. The subtypes that belong to the general category of comparison argumentation include, for instance, normative analogy (argumentation based on a model, argumentation based on the rule of justice) and descriptive analogy. Symptomatic argumentation manifests itself, among others, in subtypes such as genus-species argumentation, classification, whole-part argumentation, argumentation based on criteria and argumentation by/from authority. Just as in the case of pragmatic argumentation, the basic critical question going with the general category of argument scheme concerned must in all these cases be complemented with an additional critical question in which the specific justificatory point of the argumentation advanced in the subtype is put to the test.7 By specifying the critical questions associated with the subtypes concerned, the dialectical routes instigated by the use 7 Although

the general pragmatic principle on which the argument schemes are based remains the same for all subtypes belonging to a certain category of argument schemes, for some subtypes the basic critical question needs to be reformulated in a slightly different way. A case in point is the subtype of symptomatic argumentation based on evaluative criteria, in which a value judgment is defended by pointing at certain characteristics. The basic critical question, “Is what is claimed in the standpoint a sign of what is stated in the reason advanced?”, should then be reformulated as “Is the judgment given in the standpoint implied by the characteristic mentioned in the argument?”.

20

F. H. van Eemeren and B. Garssen

of particular subtypes of argumentation can be succinctly described in dialectical profiles. When evaluating argumentation of a specific type or subtype exactly which critical questions are pertinent in carrying out the testing procedure in a particular case depends also on the institutional preconditions of the macro-context in which the argumentation takes place. The specific conventions of the various communicative activity types that have been established in a certain domain determine to some extent which of the relevant critical questions are pertinent in a particular case and what shape they should take. This means that in dealing in the testing procedure with the argumentative moves that are made in justifying a standpoint by means of argumentation the general soundness criteria pertaining to the (sub)type of argumentation concerned that are expressed in the critical questions need to be specified or otherwise amended or complemented in accordance with the requirements of the macro-context concerned. This contextualisation is in fact what should also always happen in applying the general standards involved in the rules of the code of conduct for reasonable argumentative discourse in evaluating any of the other argumentative moves that are made in argumentative discourse. In all cases it is to be considered in the philosophical component of the research program to which extent the reasonableness of the argumentative discourse is affected by deviations from the model of a critical discussion instigated by the institutional preconditions of the communicative activity type—and perhaps also by ideological or cultural preconditions.8 As a consequence of contextual differentiation, the soundness criteria for judging the use of a specific (sub)type of argumentation may differ to some extent depending on the institutional preconditions pertaining to the macro-context in which the argumentation is advanced. Imagine two people who are playing a game of scrabble. At a certain moment one of them claims to have compiled a long word, but the other one doubts that the combination of letters that has been laid out really constitutes an English word. Now the first player uses an argument from authority to defend the claim: “This is an English word, because it is in the dictionary”. Whether the appeal to authority is in this case a sound strategic manoeuvre, depends in the first place on the kind of agreement that exists between the players on how it is to be decided whether or not a combination of letters does indeed count as an English word. The verdict on the soundness or fallaciousness of an argument from authority relates in this sense always to the starting point regarding how authoritativeness is to be decided that is operative in the macro-context in which the argumentative exchange takes place. If the players have agreed at the start of their game that a combination of letters will be regarded as an English word if it is included in the dictionary, then there is nothing wrong with the first player’s authority argument—his argumentative move The argumentation in “This book is wonderful because it presents a vivid picture of the miseries op growing up” is to be questioned by “Are books that present a vivid picture of something wonderful?”. 8 In the different kinds of argumentative practices in the various communicative activity types the “extrinsic” constraints on argumentative discourse may be to some extent determined by institutional as well as ideological or “cultural” preconditions or by a mixture of both.

2 Argument Schemes: Extending the Pragma-Dialectical Approach

21

cannot be considered fallacious and is even likely to be effective. However, the same argumentative move would be fallacious if the game was played in a macro-context in which it has been agreed upon from the start that the Concise Oxford Dictionary will be the ultimate judge while in the argumentation the arguer is referring to Webster’s. The argumentative move would be sound again if the manufacturer of the scrabble game had imposed a binding procedure for deciding about the Englishness of a word upon the players that prescribes going by a dictionary without giving any further specification as to which dictionary. If, however, the players agreed at the start of their game that a combination of letters will only be recognized as an English word if they all know the word, then the appeal to the authority of any kind of dictionary would be irrelevant and therefore fallacious. The various scenarios sketched in the scrabble example can be viewed as constituting specific macro-contexts that represent different communicative activity types or different versions of a particular communicative activity type. In specifying who or what counts as an authority, the general soundness criterion pertaining to the use of the Argument Scheme Rule involving relying on a qualified authority is in each of them implemented in a different way. In the empirical counterpart of the opening stage of the exchange a crucial starting point concerning how the game is to be decided is in each case given its own specification. In the first case, it is defined by the parties by explicitly agreeing before the argument from authority is used that the dictionary should be the specific soundness criterion that is authoritative in judging the Englishness of a word. In the second case, this specific soundness criterion is defined even more precisely by agreeing, in addition, explicitly that it is the Concise Oxford Dictionary that is to be authoritative. In the third case, the soundness criterion is defined in the same way as in the first case, but this time this criterion is simply imposed on the players as a starting point for their exchange—in the empirical counterpart of the opening stage they only have to acknowledge what the criterion involves. In the case the participants explicitly agree on a starting point that boils down to only accepting a word as English if its Englishness is recognized by all participants this starting point changes the game more drastically. In weakly conventionalized informal communicative activity types, such as a chat between friends, the specific soundness criteria applying to the argumentative moves that are made are often simply determined by the parties on the spot, when they are needed. However, these soundness criteria may also have been made familiar to the arguers in their primary socialization at home or at school, while they were growing up. In strongly conventionalized formal communicative activity types, such as a civil lawsuit, various crucial starting points, including certain evaluation procedures, are as a rule already partly or wholly given before the argumentative exchange takes place. Usually they have been explicitly taught to the participants in their secondary socialization, during their professional training as future lawyers or in some other specialised form of education. This institutional imposition of starting points, which happens particularly in strongly conventionalized and formalized communicative activity types, resembles in fact closely the third scenario just sketched. In practical terms the situation is in that case similar as in the case of exchanges with starting points based on an already existing agreement between the parties.

22

F. H. van Eemeren and B. Garssen

In the strategic manoeuvring taking place in the various kinds of argumentative practices of argumentative reality the various types and subtypes of argumentation may manifest themselves in specific, context-related ways. The ways in which a subtype of argumentation manifests itself can then be viewed as different contextuallydetermined variants of the subtype concerned. In describing the various manifestations of argumentative reality in the empirical component of the research program distinguishing between these (and other) variants is an important task.9 In this endeavour more precise distinctions can be made between variants that differ primarily in the kind of selection of the topical potential that is made (relating to differences in subject-matter), variants that differ first of all in the way they appeal to the audience (relating to different ways of associating with the listeners or readers), and variants that differ first of all in the choice of presentational devices (relating to differences in the means of expression). The general aim of this empirical research is to identify institutionally or otherwise determined variants of (sub)types of argumentation and provide accurate descriptions of the distinctive features of their sound and fallacious manifestations.

References Doury, M. (2018). Interpreting argumentation. The insider and outsider points of view. In Keynote speech at the 9th international conference on argumentation of the International Society for the Study of Argumentation (ISSA) at the University of Amsterdam. Amsterdam, July 4. Garssen, B. J. (1997). Argumentatieschema’s in pragma-dialectisch perspectief. Een theoretisch en empirisch onderzoek (Doctoral dissertation University of Amsterdam). Dordrecht: ICG printing. Garssen, B. J. (2001). Argument schemes. In F. H. van Eemeren (Ed.), Crucial concepts in argumentation theory (pp. 81–99). Amsterdam: Amsterdam University Press. Kamlah, W., & Lorenzen, P. (1984). Logical propaedeutic. Pre-school of reasonable discourse. (transl. by H. Robinsion of Logische Propädeutik. Vorschule des vernünftigen Redens. 2nd improved and enlarged ed. 1973; 1st ed. 1967). Lanham, MD: University Press of America. Lumer, C. (2011). Argument schemes. An epistemological approach. In: F. Zenker (Ed.), Argumentation. Cognition and community: Proceedings of the 9th international conference of the Ontario Society for the Study of Argumentation (OSSA). Windsor, ON: University of Windsor, May 18–22 2011. (CD-rom, ISBN 978-0-920233-66-5). van Eemeren, F. H. (2018). Argumentation theory: A pragma-dialectical perspective. Cham, Switzerland: Springer. van Eemeren, F. H., & Garssen, B. (2014). Argumentation by analogy in stereotypical argumentative patterns. In H. Jales Ribeiro (Ed.), Systematic approaches to argument by analogy (pp. 41–56). Dordrecht: Springer. Argumentation Library 25. van Eemeren, F. H., & Grootendorst, R. (1992). Argumentation, communication, and Fallacies. A pragma-dialectical perspective. Hillsdale, NJ: Lawrence Erlbaum. van Eemeren, F. H., & Grootendorst, R. (2004). A systematic theory of argumentation. The pragmadialectical approach. Cambridge, UK: Cambridge University Press. 9 When

discussing the pragma-dialectical approach to the fallacies, van Eemeren and Grootendorst (1992) already pointed out that, next to establishing general standards and specific criteria to check whether these standards have been complied with, typical manifestations of (sound or fallacious) argumentative moves need to be traced.

2 Argument Schemes: Extending the Pragma-Dialectical Approach

23

van Eemeren, F. H., Grootendorst, R., & Kruiger, T. (1978). Argumentatietheorie [Argumentation theory]. Utrecht: Het Spectrum. Aula-boeken 613. van Eemeren, F. H., Grootendorst, R., & Kruiger, T. (1983). Argumenteren [Arguing]. Groningen: Wolters-Noordhoff. van Eemeren, F. H., & Kruiger, T. (1987). Identifying argumentation schemes. In F. H. van Eemeren, R. Grootendorst, J. A. Blair, & Ch. A. Willard (Eds.), Argumentation: Perspectives and approaches. Proceedings of the Conference on Argumentation 1986 (pp. 70–81). Dordrecht: Foris. Wagemans, J. H. M. (2016). Constructing a periodic table of arguments. In P. Bondy & L. Benacquista (Eds.), Argumentation, objectivity, and bias. Proceedings of the 11th international conference of the Ontario Society for the Study of Argumentation (OSSA) (pp. 1–12). Windsor, CA: OSSA, May 18–21, 2016. Walton, D. N., Reed, C., & Macagno, F. (2008). Argumentation schemes. Cambridge: Cambridge University Press.

Chapter 3

In Search of a Workable Auxiliary Condition for Authority Arguments Hans V. Hansen

One of the methods of informal logic is the method of argument schemes and one of the most studied of the schemes is the one for Argument from Authority, also called Appeal to Expertise. The kind of authority in question is knowledge or cognitive or intellectual authority, not command or political authority. I will, for brevity’s sake, call these AA-arguments as we continue. There are two ways to approach argument schemes, the descriptive and the normative way. Descriptive argument schemes only include the necessary and sufficient conditions for an argument to belong to an argument kind, without characterizing substitution instances as either good or bad arguments. Logicians who are interested in evaluating arguments but who hold to the descriptive approach to schemes, will supplement the schemes with critical questions to guide their evaluations of arguments that satisfy the scheme (e.g., Walton, 2006). Normative schemes, however, include the conditions for being good arguments of the various kinds by building the good-making features right into the schemes. The following normative scheme for AA-arguments illustrates this point. Sc1. 1. S states that p. 2. p belongs to F. 3. F is a subject area in which knowledge is possible. 4. S has credentials that are relevant to determining whether p. 5. S’s advocacy of p is not biased. 6. There is wide agreement among relevant experts in F that p is the case. Earlier versions of this essay were presented in 2018 at CRRAR in the University of Windsor, at the Wake Forest University Argumentation Conference, at the Russell Philosophy Conference in Healdsburg, California, and at the ISSA Conference in Amsterdam. I am grateful for critical feedback received on each of those occasions. For help with this most recent version I thank CRRAR colleagues, J. A. Blair and Pierre Boulos. H. V. Hansen (B) Department of Philosophy, University of Windsor, Ontario, Canada e-mail: [email protected] © Springer Nature Switzerland AG 2020 F. H. van Eemeren and B. Garssen (eds.), From Argument Schemes to Argumentative Relations in the Wild, Argumentation Library 35, https://doi.org/10.1007/978-3-030-28367-4_3

25

26

H. V. Hansen

7. So, p should be accepted.1 According to Scheme 1, good AA-arguments are six-premise arguments with each of the premises stating a necessary condition for good arguments of this kind. The conditions are also jointly sufficient for a good AA-argument, given the proviso that this is a defeasible kind of argument.

3.1 Corroboration What I want to focus on is premise 6. This premise requires that there should be wide agreement among the relevant experts in F. I will call these “relevant experts in F” the auxiliary experts because they play a supporting role to the main speaker, or main source, S. Premises that assign a role to auxiliary experts will be referred to as “auxiliary conditions” or “auxiliary premise”. Many other informal logic and critical thinking textbooks include auxiliary premises in their models of good AA-arguments. Here is how some others have stated this requirement: • p is consistent with what other experts assert (Walton, 2006, p. 88). • There is consensus among the relevant experts supporting the claim (Bailin & Battersby, 2016, p. 137). • The experts in F agree about p (Govier, 1997, p. 140). • There should be a limited amount of disagreement over the proposition among experts in the field. The greater the degree of controversy among more-or-less equally qualified experts, the less reliable the judgment of any particular expert— (Churchill 1990, p. 471). (Italics have been added in the passages immediately above.) Two things that all these auxiliary propositions have in common are that (i) they require that all (or nearly all) of the auxiliary experts agree that p is the case, and (ii) this agreement is taken to be additional evidence for, p, over and above what the main source, S, affords. (S, the main source, is not among the auxiliary experts.) This turns the auxiliary premise into the requirement that the auxiliary experts should corroborate what S has asserted. So the auxiliary premise is to be understood as a corroboration condition (Corr-Con). We can state it explicitly as follows, (Corr-con.) The role of the auxiliary experts is to corroborate what S, the main source, has claimed.

Notice that the evidence provided by the auxiliary experts gives direct support for the conclusion of an AA-argument in addition to, and independently of, the evidence given by S, him or herself. There are problems with the Corr-con interpretation, however, that are brought out by the following dilemma. Let Γ be any AA-argument patterned on Scheme 1, 1 Adapted

from Groarke and Tindale 2013, p. 317. The premises keep the same names (numbers) going forward.

3 In Search of a Workable Auxiliary Condition for Authority …

27

S the authority relied upon, and Corr-con the interpretation given to the auxiliary premise, then we have: D1. 1. Either the auxiliary experts in Γ corroborate what S says, or they do not corroborate it. 2. If the auxiliary experts in Γ do not corroborate what S says, then Γ is not a strong argument. 3. If the auxiliary experts in Γ do corroborate what S says, then S’s testimony is of negligible value. 4. Therefore, either Γ is not a strong argument or the testimony of authority S in Γ is of negligible value. If, on the Corr-con interpretation, an auxiliary premise is not acceptable, then the AA-argument in which it occurs will not be a strong one and so we can have little faith in its conclusion. But if, on this interpretation, the premise is acceptable, then there is wide agreement among all (or most of) the relevant auxiliary experts that p is the case. Should that be the case, as it must be in a good AA-argument, then we have little need of what the source, S, might have to say2 : for now we have an even better argument for the conclusion, p—an argument from consensus. “In arguments of this type,” writes Wesley Salmon (1973, p. 93), “a large group of people, instead of a single individual, is taken as an authority. Sometimes it is the whole of humanity, sometimes a more restricted group”. If we consider the auxiliary experts as a “restricted group” whom we are consulting, then we see that Scheme 1 arguments contain all the ingredients needed for an argument from consensus. Sc2. 6. There is wide agreement among relevant [auxiliary] experts in F that p is the case. 3. F is a subject area in which knowledge is possible. 2. p belongs to F. 7. So, p should be accepted. These premises are all part of Scheme 1, but they do not at all mention S, or that S made a claim. So, with the Corr-Con interpretation of the auxiliary premise, AA-arguments turn out to be either weak arguments or they are transformed into arguments from consensus. Why do I think this is so? Because of what we may call the “comparative weight argument”. It goes like this: In normal times (not scientific revolutions), should the community of experts disagree with an authority our epistemic loyalties should be with the community of experts because the comparative weight of their collective voice counts for more than the weight of S’s solitary voice. The community of experts here is in effect the auxiliary experts of Scheme 1. Hence, when there is agreement between the authority and the experts, the view of the auxiliary experts should also carry more weight than that of the main expert alone. This argument greatly diminishes—if it doesn’t eliminate—the role, of the authority, S, in good AAarguments, and it raises the suspicion that Scheme 1 has not succeeded in capturing the conditions for a good AA-argument. 2 The

number of auxiliary experts, we suppose, is not small.

28

H. V. Hansen

Perhaps I am being uncharitable in my criticism of the AA-scheme. In those cases where there is consensus among the experts, it may be objected, we should understand that S, the main source, becomes the collection of all the experts. We should realize, it may be argued, that arguments from consensus are just a special case of appeals to authority. But what then shall we do with auxiliary premises? In those cases, comes the answer, the auxiliary premise is combined with the assertion of the main expert. But now, that is a change of scheme, isn’t it? This discussion leads directly to another reason why the Corr-Con interpretation of the auxiliary premise is problematic. Argument schemes, whether they are descriptive or normative, are schemes of individual kinds of arguments. Scheme 1 is presented as a model of good AA-arguments. The premises of individual kinds of arguments are linked together and schemes must be drawn to include this fact. If a purported scheme does not show an argument kind as having linked premise then the scheme would be of two or more argument kinds, perhaps convergently related.3 Thus, any argument scheme of a unique kind of argument must be such that any instantiation of that scheme will result in an argument with linked premises. (This demand we can call the principle of scheme integrity.) If a premise set, L, is linked, then any proper subset of L does not support the argument’s conclusion nearly as well as does the complete set of premises in L. This is one of the tests to determine whether premises are linked (Walton, 2006, p. 151). Yet we have seen that consensus arguments (Scheme2), being instantiations of a proper subset of the premises of Scheme 1, supports the conclusion as well or nearly as well, as AA-arguments that instantiate all the premises in Scheme 1. From Scheme 1 we can also extract another scheme, Scheme 3, which consists of Scheme 1 minus premise 6. Scheme 3 gives us what we may call “the basic argument from authority”4 and it, like Scheme 2, is an independent scheme. We can place Scheme 2 and Scheme 3 beside each other to form a convergent argument (for the same conclusion). Thus, Scheme 1 does not satisfy the principle of scheme integrity. This suggests that we should view Scheme 1 as an argument structure that is complex, composed of basic argument schemes. Such a configuration we can call a “non-basic argument scheme”. We should be aware that Scheme 1, on the Corr-con interpretation we have been discussing, is a non-basic argument scheme. There are then two problems that arise when we give auxiliary premises in Scheme 1 the Corr-Con interpretation. The one is that it puts the role of S, the main source, in question, thereby defeating the very purpose of an appeal to the authority of individuals or (small) groups. The other problem has consequences for scheme theory. AA-schemes purport to be basic schemes but on the Corr-con interpretation it is demonstrable that they are non-basic schemes.

3 Or

the pattern could include an idle premise. a “basic scheme” I mean one that has no independent schemes as its parts. Further insistence to get to the bare bones of appeals to authority may lead us to leave premise 5 out of Scheme 3 as well. This implies that Scheme 3 is not really a basic scheme after all.

4 By

3 In Search of a Workable Auxiliary Condition for Authority …

29

3.2 Trust Perhaps we have been barking up the wrong tree. The problem for laypeople is to put themselves in an epistemic situation wherein they can trust the testimony of experts. Jean Goodwin, following up on observations by Doug Walton, advises laypeople assessing an expert for credibility to examine such things as his employment, his degrees, the testimony of his peers, and his publications. Note that in each of these the layperson is relying on third parties, those who are likely to share the relevant expertise themselves and thus are capable of assessing the purported expert’s real knowledge. These third parties act as gatekeepers, presumably only hiring, publishing, granting degrees or good reputations to those who are truly qualified. (Goodwin 2010, p. 140; my emphasis.)

I have highlighted Goodwin’s point that there is an essential role for third parties: they are to assist laypersons in deciding on the credibility of experts. These third parties have the role of auxiliary experts. In more detail than Goodwin, John Hardwig has argued that scientific knowledge depends on trust. There are two basic reasons for this: (1) gathering and analysing data takes too long to be done by only one person, and (2) no one knows enough to be able to do all the needed experiments by herself, so research must be done by teams in which every member must trust the testimonial evidence of other team members. Although Hardwig’s main concern is to show that scientists must trust each other in order to advance their research, he does allow that “the principle of testimony” may suffice to connect lay persons with scientific knowledge: “If A [a lay person] has good reasons to believe that B [an expert] has good reasons to believe p, then A has good reasons to believe p” (Hardwig, 1991, p. 697). Here A’s reasons for believing p are at least partly what we may call “trust reasons.” However, trust reasons to believe a proposition depend on the trustworthiness of the one trusted. When we rely on what someone conveys to us (their testimony) it is because we trust their character. Hardwig considers character to have both a moral and an epistemic aspect: the moral aspect relates to truthfulness, and the epistemic aspect relates to the testifier’s reliability as an expert, to their competence, conscientiousness and epistemic self-assessment (Hardwig, 1991, p. 700). But how can we trust someone’s character if we don’t know them? Not only are we lay people far removed from experts in terms of knowledge and understanding, we are also, usually, not personally acquainted with them; hence, we are not able to place our trust in an authority or expert directly. Hardwig sees two possible solutions to this practical impediment to trust. The one is to pursue corroboration: if other scientist have come to the same conclusion, that will increases our confidence (trust) in the main source’s pronouncement. But we have already examined that proposal and found in wanting in connection with Scheme 1. The preferred solution, according to Hardwig, is to turn to intermediate sources who, although they may not have corroborated S’s findings can assess S’s trustworthiness based on their knowledge of his or her moral and epistemic character. Although Hardwig’s principle of testimony was originally stated as an immediate relationship between two parties, a giver and

30

H. V. Hansen

a receiver of testimony, he further suggests that the principle can be extended to include third parties (Hardwig, 1991, p. 701) and we can adapt it to our purposes as follows: If A [a lay person] has good reasons for believing C, D, E, … [the auxiliary experts], and C, D, E, … have good reasons for believing S [the main expert] has good reasons for believing p, then A has good reasons for believing p.

Let this be “the extended principle of testimony”. Here we have C, D, E, … in an intermediate role between A, a lay person and S, a main source. S’s reasons are presumable not trust reasons but direct evidence reasons based on her research. C, D, E, …’s reasons are, however, trust reasons based on the moral and epistemic character of S. (We may add that A’s reasons are trust reasons based on what A takes to be the trustworthiness of C, D, E, ….) This leads us to an alternative to the Corcon interpretation for auxiliary premises. We now focus on having evidence for the credibility of S, the main source. We can call it the Credibility Condition (Cred-con) interpretation of the auxiliary premise. More precisely, (Cred-con.) The only role of the auxiliary experts is to confirm that S, the main source, is credible with regard to the issue in question.

What the auxiliary experts say may indirectly support the conclusion of an AAargument by bolstering the credibility of S, the main source. Cred-con does not imply that the auxiliary experts know and/or understand S’s reasons for p, the conclusion in question; yet they may still have confidence in S’s judgment based on their knowledge of S’s past accomplishments and her reputation. Hence, the role of auxiliary experts on the Cred-con interpretation is quite different than it is on the Corr-con interpretation because it shifts the focus away from looking for further support for the conclusion, p, to looking for support for the trustworthiness of S. I think that this modification ameliorates the problems we were led to by Corr-con. Let us restate the dilemma, D1, under the new interpretation, Cred-con. D2. 1. Either the auxiliary experts in Γ find S trustworthy or they do not find S trustworthy; 2. If the auxiliary experts in Γ do not find S trustworthy, then Γ is not a strong argument; 3. If the auxiliary experts in Γ do find S trustworthy, then S’s testimony in Γ is credible (other conditions being satisfied); 4. Therefore, either Γ is not a strong argument or S’s testimony in Γ is credible (other conditions being satisfied). The outcome of this second dilemma based on the Cred-con interpretation is more favourable to the role of S, the main source in AA-arguments, than was the outcome based on the Corr-con interpretation. True, it is still the case that if the auxiliary experts in an AA-argument do not find S credible then the argument will be a weak one. However, should the auxiliaries find S trustworthy this will give S’s testimony a boost of support and make the AA-argument stronger. Thus, on the Cred-con reading

3 In Search of a Workable Auxiliary Condition for Authority …

31

the auxiliary experts bolster S’s testimony rather than marginalize its value, as they did on the Corr-con interpretation. But now we are faced with a “book-keeping” problem, if we want to retain the other insights of Scheme 1. The fourth premise of the scheme already requires that S has credentials, and having credentials—degrees, titles, awards – is a mark of recognition by one’s peers. This makes for an unnecessary overlap of requirements expressed in Premise 4 and the Cred-con reading of the auxiliary premise. In fact it leads us to conclude that, on the Cred-con interpretation, if the auxiliary premise is acceptable, then Premise 2 is superfluous. Is there a way around this? Our response is to distinguish the various factors that contribute to a source’s credibility. One source of confidence in S will be that she has the accreditation to practice in her field. This is evidenced by degrees, diplomas and memberships in professional organizations and, as we just observed, is demanded by premise 4. A lay person can, in many cases, explore and decide for himself whether S satisfies the accreditation condition; e.g., I can easily find out from where Noam Chomsky got his degrees. It is less accessible to a layperson whether S has any biases or interests that might affect her research and pronouncements, a concern brought forward by Premise 5. This is something professional colleagues, the auxiliary experts, are more likely to know about. A third source of credibility has to do with what Hardwig called “epistemic character”. Here we are thinking of what contributes to S’s reputation among her peers. How does one progress from being a qualified practitioner with the necessary degrees to being a recognized authority who has the confidence of her professional colleagues and can be cited in an AA-argument? This is where we consider questions such as: Has S successfully published her research in peerrecognized journals and with respected publishers? Are her conclusions, in general, thought to be reliable? Has she in the past responded appropriately to criticisms? Does she fulfill her commitments? In short, what is S’s track record as a member of the research community? Considerations such as these are the measure of a researcher’s epistemic character, and it is this source of trust that lay persons will have considerable difficulty in answering without the assistance of auxiliary experts. Our auxiliary premise may now be revised and stated as follows: There is wide agreement among relevant [auxiliary] experts in F that S has good epistemic character. This epistemic character condition has some overlap with premise 5 (that S must not be biased), and so redundancy seems to threaten us again. But here we can mark an important distinction between those biases due to character and those due to the community standards or to faulty equipment or research design. These are different kinds of biases: the former belong to epistemic character, the latter properly belong with premise 5. There are, then, three distinct factors that contribute to the credibility of a source’s pronouncements: accreditation, epistemic character, and lack of biases. The factors may be seen as logically independent of each other, any one of them could fail to obtain while the other two do not. It follows that the more of them do obtain, and the extent to which each of them is satisfied, the greater should be our confidence in

32

H. V. Hansen

the sources’s credibility. For clarity’s sake, it is desirable to group these credibilityrelated factors together in AA-arguments. This gives us a three-premise convergent argument pattern: Ptn1. 4. S has credentials that are relevant to determining whether p. 5. S’s advocacy of p is not biased. 6. There is wide agreement among relevant [auxiliary] experts in F that S has a good epistemic character. (Cred-con) 8. So, S’s pronouncements about propositions in F are trustworthy. The conclusion of this argument pattern introduces a new proposition into the scheme for an AA-argument, namely proposition (8), that S’s pronouncements about propositions in F are trustworthy. My suggestion is that we restructure the scheme for AA-arguments such that it consist of four linked premises as follows: Sc4. 1. S states that p. 2. p belongs to F. 3. F is a subject area in which knowledge is possible. 8. S’s pronouncements about propositions in F are trustworthy. 7. So, p should be accepted. These premises are linked together because no proper subset of them will support the conclusion nearly as well as the full set of premises. (That is the test for linkage.) So, scheme 4 is a basic argument scheme. But now, what has happened to the auxiliary experts? How do we bring their role to light? They are relegated to providing support for a sub-argument for one of the main premises in scheme 4. With the argument Pattern 1 included, altogether we have: Ptn2. 1. S says that p. 2. p belongs to F. 3. F is a subject area in which knowledge is possible. 4. S has credentials that are relevant to determining whether p. 5. S’s advocacy of p is not biased. 6 . (Cred-con). There is wide agreement among relevant [auxiliary] experts in F that S has a good epistemic character. 8. So, S’s pronouncements about propositions in F are trustworthy. (4–6) 7. So, p should be accepted. (1, 2, 3, 8). In this pattern of argument (4–6 ) convergently support (8) and (1, 2, 3, 8) are linked with respect to the conclusion (7). Now, it may seem that Pattern 2 is subject to the same objection I brought against Scheme 1: that it is really a conglomeration of two basic schemes converging on the same conclusion. But this objection is not right for two reasons: (i) Pattern 1 and Pattern 2 do not have the same conclusion as convergent arguments do, and (ii) Pattern 1 is not an argument scheme. Pattern 1 and Scheme 4 are related not as convergent arguments but as serial arguments (Walton, 2006, p. 146), with the conclusion of Pattern 1 supporting a premise in Scheme 4. Pattern 2, which serially combines Pattern 1 and Scheme 4, is not a basic scheme for the reason that no basic schemes include premises that evidentially support another

3 In Search of a Workable Auxiliary Condition for Authority …

33

premise of the scheme; hence, the array of propositions that constitute Pattern 2 is not a scheme. The advantages of seeing the role of the auxiliary experts as testifying to a source’s epistemic character rather than being corroborators of the conclusion has two advantages. First, it avoids the result that an AA-argument’s main source is marginalised by a consensus argument. Second, it avoids the result that AA-arguments, as modelled on Scheme 1, turn out to be patterns of independent convergent support rather than a pattern of evidentially dependent propositions as in Pattern 2. Nevertheless, there are reasons to be discontent with our conclusion thus far.

3.3 The Layperson’s Paradox The central problem of appeals to authority is to bridge the epistemic gap between laypersons and experts. Overcoming the problem seems doomed from the outset. Goldman expresses the problem this way: If the layperson is just that, a layperson, how can he or she hope to determine whether past beliefs of each expert … [are] … true or false, and hence whether they have good or bad past track records? It’s the inability of the layperson to settle the truth-values of such propositions in the first place that launched his/her turning to experts! How can the layperson now hope to assess the experts by turning to the sort of facts of which he despaired from the beginning? It is the very esotericness of such facts that gets the problem rolling in the first place. (Goldman, 2011, p. 216)

There is here what we may call the Layperson’s Paradox: In order for laypersons to avail themselves of the knowledge that authorities/experts have, laypersons must themselves have the kind of knowledge that the authorities/experts have. How can this paradox be dispelled? The simplest solution is for laypeople to get themselves educated in the fields in which they desire to have knowledge: make themselves peers of the expert whose opinion concerns them, then they will be able to evaluate her assertions on their own without anybody’s help. That way of dispelling the paradox, we are assuming, is off the table. Our attempted solutions so far have been to give an intermediary role to auxiliary experts: they are posited as the middle-people between us and the main experts. On one interpretation the auxiliary experts were cast as corroborators of what source experts have claimed; another interpretation gave them the role of vouching for the trustworthiness of a main source. However, both of these roles focussed on the way auxiliary experts can support (or not) the main experts. The question that is of equal, if not more, importance is how they come to the aid of the laypersons who seek their help in evaluating what an expert source has asserted. The auxiliary experts are themselves experts and they are as removed from laypeople as are the main experts. We must ask: (a) how do we identify the auxiliary experts? And (b) how can we understand and trust them? And (c) must we really consult all or most of them as the auxiliary condition in Schemes 1, 4 and 6, seem to require? The answers to these questions are as follows: (a’) Who the auxiliary experts are in

34

H. V. Hansen

a given field is determined the same way that we decide whether a main expert is an expert in a field: credentials, track record, etc. Thus the same epistemic distance exists between laypeople and auxiliary experts as exists between laypeople and main source experts. (b’) We must examine their assertions with the same care with which we intended to examine the statements of the main expert. Do we then need another set of auxiliary experts to help us evaluate what the first set of auxiliary experts say about the main source? The answers to the first two questions shows us that the problem of expertise (the layman’s paradox) has not gone away by our relying on auxiliary experts. And, as for the last question, (c’), it is practically impossible for a layperson to consult the views of all or most of the auxiliary experts in any field of knowledge (even if he could identify them) in order to determine whether or not there is wide agreement among them. With these three answers in mind we are led to think that the idea of auxiliary experts is not as helpful as we thought it might be after all, because they do not help us solve the practical problems facing laypersons. The auxiliary experts are, rather, obstacles. What means are available, other than the inspection of credentials and track records, to give laypersons confidence in experts’ opinions, if we disband the auxiliary experts? Goldman (2011, pp. 117–18) offers us at least two suggestions, the first introduces a distinction between direct and indirect argumentative justification. If we test an expert’s assertion by examining the evidence in his premises and the inferences he draw from his evidence, this is testing his views by direct argumentative justification. It is a means laypersons will find it nearly impossible to pursue because experts’ arguments depend on the kind of knowledge laypersons do not have. However, if we test an expert’s competence indirectly by observing how he fares in debates with other experts in his field, we may be able to observe his dialectical proficiency. By this method the layperson is not required to deal directly with specialized knowledge but only with the experts’ argumentative performances. Even though our layperson knows nothing about socio-biology, for example, she may come to believe that one of two debaters is more competent than the other in the subject by his readiness to answer questions, introduce rebutters, point out inconsistencies in his opponent’s position, and include additional relevant evidence. The debater who is most proficient in this exercise demonstrates “dialectical superiority” and this, by an inference to the best explanation, “may reasonably be taken as an indicator of [his/her] having superior expertise on the question at issue” (Goldman, 2011, p. 118). We then make the further inference that the expert who is dialectically superior is more likely to be right. Goldman’s other suggestion, designed as a layperson’s check on an expert’s track record, relies on the esoteric-exoteric knowledge distinction. Expert knowledge, consisting mainly of esoteric propositions (not understandable by laypersons), may also contain some exoteric propositions (understandable by laypersons), or have exoteric consequences. Goldman’s example is of an expert predicting now (in 2019) that there will be a solar eclipse in Santiago, Chile, in 2025. From my present point of view, as a lay person, this is a claim based on esoteric knowledge and I cannot evaluate it. However, it becomes exoteric for me in six years, especially if I am lucky enough

3 In Search of a Workable Auxiliary Condition for Authority …

35

to be in Santiago. Depending on what happens in the sky that day, my confidence in the expert will increase or decrease (Goldman, 2015, pp. 216–217). Interesting as Goldman’s suggestions are, he realizes that they are methods with only limited utility (Goldman, 2011, p. 129). Seldom do we have occasion to compare the dialectical competence of debating experts, and rarely do we have the opportunity to test an expert’s competence by waiting for the exoteric outcome of one of her predictions. We are in search of better ways, something approaching a series of steps laypersons can take in their efforts to evaluate the claims of experts.

3.4 Goldberg Variations Let us consider our problem anew, this time viewing it from the perspective of the epistemology of testimony. According to Sanford Goldberg there are two kinds of testimonial evidence, one of which is testimony-to-truth. Evidence of this kind includes the transmission of a proposition, p, from a source to a hearer through testimony. The main source, S in an AA-argument, by her assertion, p, offers testimony of this kind. This, too, is what auxiliary experts will do, if they have a role to play. The other kind of evidence is not itself testimony but is related to it. It consists in the absence of testimony where we expected to find it. In failing to testify (report) that p, as we anticipated it would do, a source can lend support to not-p. This second kind of “testimonial” evidence Goldberg describes as exhibiting a truth-to-testimony relationship, i.e., because it is not true that p, there is no testimony that p. This suggests the possibility that it may be more propitious for laypeople to look for truth-to-testimony evidence that would give them reasons to reject or doubt an expert’s testimony than for additional support of his views? Let us open this door a little further. Goldberg’s paper is aptly named, “If that were true I would have heard about it by now”. To show how non-testimony can have evidential weight he develops the idea of coverage-supported beliefs which must satisfy the following conditions. First, there must be an organization in the layperson’s epistemic community that studies questions arising in the field to which p (whatever the proposition in question is) belongs. For example, if we are wondering whether the level of pollution in Lake Huron has increased in the last five years (H), there must exist an organization, O, that investigates water quality in the Great Lakes. Goldberg calls that the existence condition. The second condition is that O reliably reports its findings. If the level of pollution has been increasing in Lake Huron then a media source, M, aware of O will report it, and M’s reports are reliable. This is the coverage-reliance condition. Next is the sufficient-time condition: O must have had sufficient time to investigate the proposition H and make a determination. This means that there may well be a lag between an event or a scientific discovery and the time that M reports it. A fourth condition is that the knowledge-seeker who is interested in H is so situated in the (epistemic) community that she will have heard about it through M if H is the case. That is called the reception condition. Finally, the silence condition, is that our knowledge-seeker has not heard anything that supports H through M or

36

H. V. Hansen

other sources in her epistemic community. With these five conditions of coverage sufficiency satisfied, the belief that not-H can be inductively justified. Goldberg (2011, p. 95) suggests that the reasoning fits a modus tollens pattern like this: If p, then one or more among the trusted sources testified that p. None of the trusted sources testified that p. So, it is not the case that p.

(Substituting H for p in our imagined case about Lake Huron, we get the conclusion that the level of pollution in Lake Huron has not increased in the last five years (not-H). Goldman (2015, p. 212) explains the inference from the non-reporting of p to p as an inference to the best explanation of the silence that is “heard”. How can Goldberg’s ideas about the evidence of non-testimony be adapted for auxiliary premises in AA-arguments? The principal aim of informal logic is to develop methods for evaluating natural language arguments by non-formal means. It is also in the spirit of informal logic that its methods should be as widely usable as possible and not demand any special technical skills of the users, as do formal logic and probability theory. The methods of informal logic thus aim to be user-adequate, meaning that they are suitable to the knowledge and abilities of the argument users. What must be the case if auxiliary premises are to be included in AA-arguments is (i) that they contribute to the evaluation of AA-arguments, and (ii) that they are evaluable by laypersons without the help of auxiliary experts. Applying these desiderata requires that each of the premises in an AA-argument must be evaluable by laypersons directly, and not indirectly through the testimony of others. As Goldberg develops his view it is not especially tailored for laypeople or AAarguments. His purpose is to make us aware of the evidence that can come from non-testimony. But we can adapt his ideas to AA-arguments and maybe find a way out of the layman’s paradox by taking the following steps: (i) specify that the scope of coverage sufficiency for laypersons is the information outlets they can understand; and (ii) direct laypeople to be open to evidence that contradicts S’s claim that p rather than evidence that directly confirms p. This idea needs some fine-tuning. Evaluating AA-arguments is unlike evaluating most other kinds of arguments, ad hominems or slippery slopes, for example, because it requires arguers to be more pro-active, to take the initiative in getting whatever information they can to bolster their evaluation. In adapting Goldberg’s analysis of coverage sufficiency to our purpose we presume that laypersons exist in epistemic communities in which exoteric information is available; that is, information for the field in question that requires little or no specialized knowledge. There are science magazines for lay people (e.g., Scientific American, American Scientist, Psychology Today, etc.); and radio shows such as Quirks and Quarks (CBC); television programs like The Nature of Things (CBC), Horizon and Planet Earth (BBC). There are various “science feeds” available on the internet: Google News /Science, and also a great many devoted to particular regional interests. For example, for those who worry about H (the level of pollution in Lake Huron), there are websites like Canada’s Great Lakes Areas of Concern,

3 In Search of a Workable Auxiliary Condition for Authority …

37

and the American based Great Lakes Observing System.5 These outlets of digestible information are Janus-faced sources (in a good way). They have the ability to receive and understand the esoteric knowledge of experts and the willingness to turn it into exoteric information for laypeople. By responding in this way to the layman’s paradox we are attempting to move the exercise of our critical skills from the evaluation of an expert source, S, where they have little traction, to the area of popular outlets of information where our critical skills have much more bite. In including popular websites among the outlets that can be monitored for information, we at the same time require that only sources that satisfy critical standards should be used. It is part of our solution to the layman’s paradox that laypeople are better able to all evaluate whether websites can be trusted than they are able to evaluate the esoteric claims of physicists, environmentalists, medical doctors and other experts. The second step out of the layperson’s paradox is to look for information that is inconsistent with what the main source has claimed. The requirement that we should find universal agreement for a claim is one that is nearly impossible to satisfy. The alternative is to observe whether the sources being monitored publish any disconfirming evidence that stands out. If such there is we should take an agnostic attitude toward the source’s pronouncements. Conversely, if no such reports come to light then we have an additional reason to find the source’s sayings acceptable. Thus, laypersons should take a falsification, rather than a verification, approach in their inquiry concerning a source. In order to make this solution to the layman’s paradox work, there are two kinds of conditions to satisfy: those having to do with opportunity and those with epistemic responsibility. The opportunities that must exist for laypersons in evaluating expertise are as follows: 1. There must exist outlets that publish information for laypeople in field F. 2. The outlets report in a timely fashion. 3. Laypersons are likely to receive information as it becomes available from the outlets. The epistemic responsibilities that taking advantage of these opportunities demand are the following: 4. Laypersons must critically evaluate the sources they rely upon. (There are checks to run on publishers and websites) This is not unlike checking the reliability of the auxiliary experts, but it is more doable. 5. Laypersons must be open to receiving information from a number of different sites, independent of each other. With these conditions satisfied, and no significant testimony received that not-p, one has more evidence to believe what the expert asserted. It amounts to a strengthening of an AA-argument whereas contradictory evidence raises hesitations about the main source’s claim and should weaken our confidence in her pronouncements. 5I

thank Dr. R. M. L. McKay, Director of the Great Lakes Institute for Environmental Research in the University of Windsor, for bringing these websites to my attention.

38

H. V. Hansen

Let us then attempt another interpretation of the auxiliary condition, this time calling it the Coverage Condition. (Cov-con.) The purpose of the auxiliary condition is to provide additional support for main source S’s claim that p by providing evidence that there is no or little support for not-p.

The next problem is how to formulate what Cov-con demands as a premise for an AA-scheme. Here is a proposal: “No credible evidence that not-p has come to light in available sources.” Our revised scheme will be Scheme 1 with a new auxiliary condition (Sc5.6 ) replacing the original (Sc1.6). Sc5. 1. S states that p. 2. p belongs to F. 3. F is a subject area in which knowledge is possible. 4. S has credentials which are relevant to determining whether p. 5. S’s advocacy of p is not biassed. 6 . No credible evidence that not-p has come to light in available sources. 7. So, p should be accepted. Will this scheme for AA-arguments fare any better than the others we have proposed? Let us discuss it.

3.5 Discussion (1) User-Adequacy. The Cov-con interpretation of the auxiliary condition is the solution to the layman’s paradox in two ways. First, because coverage sufficiency with respect to any field, as explained above, is within the means of lay people through popular sources. True, some mystery remains about how esoteric scientific discourse gets turned into lay-speak in reliable sources, accessible to the uninitiated, but that is just something that as reliabilists we must accept as happening—this is where trust has found a new role to play. The second way that the Cov-con interpretation works to overcome the paradox is that it gives up the demand included in many versions of the auxiliary premise that all or most experts in the field must be consulted and voice their support. It is practically more efficient to look for instances of disagreement than it is to find (near) universal agreement. I suspect that many authors who have included the requirement that the relevant (auxiliary) experts (nearly) all agree with the main expert really meant that there was no known disagreement among the relevant experts. This is the gist of the passage from Churchill quoted earlier. But, to be specific, Cov-con overcomes a problem that was common to both the Corr-con and Cred-con interpretations which both imposed unrealistic expectations on argument evaluators by making it a necessary condition that they should be assured that all or most other experts were in agreement. (2) Importance of the main source. Our argument against the Corr-con interpretation was that the auxiliary experts “took over the show” and thereby marginalized the

3 In Search of a Workable Auxiliary Condition for Authority …

39

main expert. If all the relevant experts agree with the proposition p, that is worth more, went the argument, than individual expert S’s claim that p. Hence, AAarguments with the Corr-con interpretation seem to self-destruct as appeals to an authority. This does not happen on the Cov-con interpretation of the auxiliary premise. A good lack-of-evidence argument supports but does not eclipse the importance of S’s testimony. So it avoids the “comparative weight” argument. The dilemma that gave rise to our original worries comes out better on the Cov-con interpretation. D3. 1. Either the auxiliary sources in Γ find evidence for not-p or they do not find evidence for not-p. 2. If the auxiliary sources in Γ find evidence for not-p, then the credibility of S’s claim that p is weakened. 3. If the auxiliary sources in Γ do not find evidence for not-p, then the credibility of S’s claim that p has additional support. 4. Therefore, in Γ , either the credibility of S’s claim that p is weakened or it is has additional support (3) Return to corroboration. Earlier we found reasons for giving up on Corr-con, and so we turned to indirect evidence as a way to increase our confidence in the conclusion through evidence of S’s good epistemic character. This may have left us a little uneasy. Why give up on possible additional evidence for p altogether? Cov-con offers an amelioration: extra evidence is in fact what we are finding on this interpretation by the truth-to-testimony approach. The absence of contrary testimony is a kind of corroborating reason for the conclusion of the argument, but it remains a secondary source of support. (4) Scheme integrity. The Cov-con interpretation does return us, however, to the problem of scheme integrity. Scheme 5 is really a combination Scheme 3 (Scheme 1 without the auxiliary premise) and this scheme: Sc6. 2. p belongs to F. 3. F is a subject area in which knowledge is possible. 6 . No credible evidence that not-p has come to light in available sources. 7. So, p should be accepted Scheme 6 is a variation on the scheme for the argument kind “appeal to ignorance” or “lack of evidence”, or argumentum ad ignorantiam. Sometimes arguments of this ilk are called fallacies but here that charge will not stick simply because if coverage reliability is satisfied, there is no epistemic irresponsibility, although the conclusion is only inductively supported. Moreover, the appeal to the absence of evidence is made in conjunction with an argument that does give positive support for the conclusion. Schemes 3 and 6 combine convergently to form Scheme 5. The conclusion comes back to us that if AA-arguments are modelled (or analysed) on the pattern of either Scheme 1 or Scheme 6, they are not basic argument schemes.

40

H. V. Hansen

5. Whither the auxiliary experts? So, have we retired the auxiliary experts? In one way yes: because they are gone from Scheme 5. But perhaps we have reinvented them in another way. They have morphed into the laypeople accessible sources that we monitor for information that could give us reasons to doubt the acceptability of S’s claims.

References Bailin, S., & Battersby, M. (2016). Reason in the balance (2nd ed.). Indianapolis: Hackett Publishing. Churchill, R. P. (1990). Logic: An introduction. New York: St Martin’s. Goldberg, S. C. (2011). If that were true I would have heard about it by now. In A. I. Goldman & D. Whitcomb (Eds.), Social epistemology: Essential readings (pp. 92–108). New York: Oxford University Press. Goldman, A. (2011). Experts: Which ones should we trust? In A. Goldman & D. Whitcomb (Eds.), Social epistemology: Essential readings (pp. 109–113). New York: Oxford University Press. Goldman, A. (2015). Testimony and disagreement. In A. I. Goldman & M. McGrath (Eds.), Epistemology: A contemporary introduction (pp. 205–223). New York: Oxford University Press. Goodwin, J. (2010). Trust in experts. In C. Reed & C. Tindale (Eds.), Dialectics, dialogue, and argumentation (pp. 133–143). London: College Publications. Govier, T. (1997). A practical study of argument (4th ed.). Belmont, CA: Wadsworth. Groarke, L., & Tindale, C. (2013). Good reasoning matters! (5th ed.). Don Mills, ON: Oxford University Press. Hardwig, J. (1991). The role of trust in knowledge. Journal of Philosophy, 88, 693–708. Salmon, W. (1973). Logic (2nd ed.). Upper Saddle River, NJ: Prentice-Hall. Walton, D. (2006). Fundamentals of critical argumentation. New York: Cambridge University Press.

Chapter 4

Give the Standard Treatment of Fallacies a Chance! Cognitive and Rhetorical Insights into Fallacy Processing Steve Oswald and Thierry Herman

4.1 Introduction The identification of normative standards of argument quality can be said to be one of the cornerstones of argumentation scholarship, so much so that “[t]he degree to which a theory of argumentation makes it possible to give an adequate treatment of the fallacies can even be considered as a litmus test of the quality of the theory” (van Eemeren et al., 2014, p. 25). Across the ages, various normative accounts have come up with different standards meant to discriminate good from bad arguments. In his influential work on fallacies and on the history of their study, Hamblin (1970) referred to most accounts of fallacious argumentation available in his time as “the standard treatment”. Under this account, a fallacious argument is an argument that “seems to be valid but is not so” (Hamblin, 1970, p. 12). Hamblin, and others in his wake, abundantly criticised this definition, notably for its inability to identify stable and accurate norms for what should count as (un)acceptable argumentation. While we acknowledge the inadequacy of the standard treatment when it comes to assessing the normative dimension of argumentation, we suggest that its definition of fallacious arguments sheds light on another issue of interest to argumentation 1 scholars, namely rhetorical effectiveness. In this chapter, we thus presuppose the 1 By

rhetorical effectiveness, we mean any type of influence that the argument might have on the addressee and that is consistent with the speaker’s persuasive intentions. S. Oswald (B) English Department, University of Fribourg, Fribourg, Switzerland e-mail: [email protected] T. Herman Institute of Information and Language Science, University of Neuchâtel, Neuchâtel, Switzerland e-mail: [email protected] School of French as a Foreign Language, University of Lausanne, Lausanne, Switzerland © Springer Nature Switzerland AG 2020 F. H. van Eemeren and B. Garssen (eds.), From Argument Schemes to Argumentative Relations in the Wild, Argumentation Library 35, https://doi.org/10.1007/978-3-030-28367-4_4

41

42

S. Oswald and T. Herman

existence of normative standards to assess argument quality, but we wish to move away from normative concerns in order to explore the question of rhetorical effectiveness. Specifically, we discuss the fundamental intuition regarding the way fallacies are processed, and which is contained in Hamblin’s definition: if fallacies are misidentified as valid (or acceptable/reasonable/unproblematic), then it means that their fallaciousness remains undetected. This, in itself, justifies an inquiry into how they manage to remain undetected. We propose to outline the contours of this inquiry by drawing on an explanatorily adequate cognitive pragmatic account of meaning. In doing so, we will defend the claim that rhetorically effective fallacies cognitively constrain the addressee’s argumentative evaluation of the message, so as to bypass or withstand critical testing. This involves, on behalf of the speaker, the formulation of argumentative messages that are likely to strengthen the content of the fallacy to make it epistemically stronger and cognitively easy to process and/or weaken potential or actual refutations on the same two dimensions (epistemic strength and ease of processing). In a nutshell, this contribution (i) develops the rationale for envisaging rhetorical effectiveness in terms of foregrounding and backgrounding cognitive processes, and (ii) explains how the linguistic choices made in fallacious arguments are likely to decisively constrain these cognitive processes in their attempt to secure rhetorical effectiveness. Section 4.2 starts with a discussion of the standard treatment of fallacies and reports on some normative issues it has faced, before proceeding to defend a rhetorical reading of the definition of fallacy it offers in an attempt to salvage it. Section 4.3 offers a cognitive and reductionist account of how argumentative evaluation unfolds, with specific focus on fallacies. It then outlines an account of what rhetorical effectiveness is, from a cognitive perspective, focusing on parameters which determine informational selection. Section 4.4 illustrates how these mechanisms play out by discussing various backgrounding and foregrounding strategies at play in a courtroom dialogue taken from the 2009 film Public Enemies, directed by Michael Mann.

4.2 The Standard Treatment of Fallacies: Problems and Opportunities 4.2.1 Normative Issues Hamblin, in his ground-breaking study on fallacies (Hamblin, 1970) observes that most available accounts of fallacious argumentation in the logic textbooks of his time seem to have much in common. The vast majority, according to him, are very traditional (few of them go beyond Aristotle’s typology) and they all present similar typologies of finite sets of arguments, which are said to be fallacious. He dubs these accounts ‘the standard treatment of fallacies’ and levels devastating criticism against it, as, according to him, it is highly unsatisfactory and problematic on a number of counts. To quote his own words, this treatment is “as debased, worn-out and

4 Give the Standard Treatment of Fallacies a Chance …

43

dogmatic a treatment as could be imagined—incredibly tradition-bound, yet lacking in logic and historical sense alike, and almost without connection to anything else in modern logic at all” (Hamblin, 1970, p. 12). Accordingly, he considers the landscape of fallacy research to be so gloomy that it has prevented argumentation scholarship from elaborating any serious account of fallacious argumentation. Hamblin’s criticism begins with the identification of what he takes to be a common definition of fallacy within the standard treatment. Under this account, a fallacious argument is an argument that “seems to be valid but is not so” (Hamblin, 1970, p. 12, original emphasis). Since this is a widespread definition, any critique of the definition inevitably disqualifies all existing accounts which can be said to take it on board. Hamblin first notes that, while this definition should make us expect researchers to focus on either what makes fallacious arguments seem valid or on what makes them invalid, virtually no one has risen up to the challenge. Moreover, as Hamblin, but also many others after him, pointed out, adopting such a definition leads to a number of theoretical and methodological problems, among which we find the following: • the definition only seems to capture a restricted set of formal fallacies and thus fails to capture most types of fallacies (see also van Eemeren, 2001) • the definition captures arguments which we would be ready to deem fallacious, but which are valid from a logical perspective, such as petitio principii (see also van Eemeren, 2001) • the definition makes fallacy judgements relative (since, in the end, what seems valid to someone might not to someone else, see Walton, 2010) • the definition makes fallaciousness an issue of formal validity, which, again, fails to capture the reality of argumentative fallaciousness (see, again, Walton, 2010, and more generally any account of the rise of informal logic, such as Johnson & Blair, 2006, or Johnson, 2014). We leave here issues about the inadequacy of the standard treatment in any normative inquiry and refer the reader to the abovementioned literature for an informative discussion of the normative aspects of argument quality. Instead, we wish now to highlight that there are grounds, and in fact argumentatively relevant grounds, to consider that we should not throw out the definition of fallacies given in the standard treatment with the bathwater.

4.2.2 Salvaging the Standard Treatment: Rhetorical Insights We submit here that the definition of fallacy given in the standard treatment contains a basic, yet valuable, intuition about why fallacies might have some appeal to some audiences. In this sense, we propose to depart from logical or dialectical normative considerations on the standard treatment of fallacies and offer a rhetorical reading of the definition. We then draw its implications for the study of argumentation. We propose that Hamblin put his finger on the very mechanism that allows

44

S. Oswald and T. Herman

fallacious arguments to have some rhetorical appeal (which still needs to be defined, see Sect. 4.3.3). Under the standard treatment, fallacies are defined as invalid arguments that seem valid. That is, that seem valid to someone—crucially, to those who fail to identify them as fallacies. It stands to reason that subjects who recognise that, in a line of argumentation offered by a speaker, the conclusion does not follow from the premises brought to bear in the justification would in principle reject it; other things being equal, it is indeed part of one’s default epistemic goals not to take on board information that one thinks ill-evidenced. This first specification allows us to establish that one of the core features of successful fallacies has to do with the fact that there is something about them that makes them undetectable. In other words, successful fallacies are those that manage to prevent their addressees from identifying that, how and/or why they are fallacies. Therefore, a reasonable stance on the issue would be to consider that the rhetorical success of a fallacy is intimately related to the manifestness of its fallacious nature: the less manifest the fallacious status, the more undetectable, and the more rhetorically effective the fallacy. Now, if we are to further explore this line of reasoning, it seems important to be clear on what the notion of manifestness amounts to. Drawing on Sperber & Wilson’s relevance theory, we will consider that the manifestness of information is defined in terms of its ability to be represented: “a fact is manifest to an individual at a given time if and only if he is capable of representing it mentally and accepting its representation as true or probably true” (Sperber & Wilson, 1995, p. 39). In the case of fallacious argumentation, it is therefore crucial to its success that any information related to the fallacy’s status qua fallacy remains non-manifest in whatever information sets the addressee is processing as he encounters the fallacy. We will come back to this as we develop our account below to show how manifestness can be tweaked in the search for rhetorical effectiveness. Let us now further reflect on the theoretical implications of this idea for the type of inquiry we have in mind. Determining the parameters that make a fallacy undetected ultimately appears to be a psychological issue whose investigation must unfold within some sort of cognitive or psychological framework. This is because what is at stake in that line of inquiry are (i) explanations of why a given set of information fails to be identified as normatively lacking, and, thus, (ii) explanations accounting for why this set of information is not rejected—as it should, in light of its failure to uphold the normative standards in place. In other words, explaining why fallacies can go unnoticed will lead us to wonder about more general cognitive mechanisms underlying the selection (i.e., the inclusion and/or rejection) of information in communicative exchanges. While the disruptive nature of fallacies is not a new idea in argumentation scholarship, it has not, to our knowledge, been developed so far within a consistent research programme meant to account for their cognitive underpinnings. However, some relevant work paving the way for this investigation deserves to be mentioned here. In her paper on the relationship between fallacies and heuristics, Jackson (1996) notes that fallacies “are not incorrect argument schemes or correct argument schemes applied

4 Give the Standard Treatment of Fallacies a Chance …

45

incorrectly, but failed diagnostic strategies” (1996, p. 111) and that they are persuasive “because audiences use them as shortcuts to avoid careful thinking about issues” (1996, p. 104). Here Jackson (rightfully) sees the relationship between fallacies, taken as discursive argumentative objects, and cognition as a disruptive one. At the end of the day, fallacies function by perturbing critical thinking processes in order to generate some kind of error—usually consisting in the non-identification of the fallacy as one, and consequently leading to a failure to reject it, this resulting in a so-called “failed diagnostic strategy”. Willard (1995) holds a similar and compatible view when he states that fallacies “distract the persuadee from reflective thinking”, “divert attention from argumentatively weak messages” and “distract attention from matters the analyst takes to be relevant” (1995, pp. 148–149). These considerations resonate with the definition of fallacy offered in the standard treatment in that they seem to specify, albeit without entering into the details, what effect (successful) fallacies have on people’s minds. Essentially, they disrupt evaluative processes and cause their victims to fail to identify them as fallacies, and thus to fail to reject either their conclusion or the justificatory relationship that is offered to support it. In a related area of inquiry, Correia (Correia, 2011, 2014) develops the link between cognitive biases, motivated reasoning and fallacies—albeit from the perspective of the speaker—in order to show that many fallacies are likely to be caused by cognitive biases. We suggest that it is possible to conduct similar reflections about the reception end of the communicative process, namely the addressee, and that one of the fundamental features of fallacies is their ability to exploit the inherent fallibility of cognitive mechanisms. Let us take stock of all this. What we propose is to tackle the rhetorical effectiveness of fallacious arguments via a cognitive framework, based on the assumption that the study of rhetorical effectiveness ideally lends itself to a cognitive investigation. Moreover, we contend that the definition of fallacy given in the standard treatment contains a valuable starting point for this project: the appearance condition (as discussed by Hansen, 2015) stipulates that the condition under which fallacies are successful—whether that is good or bad is a separate question—has to do with their ability to go unnoticed. Thus, what we should explain in order to tackle rhetorical effectiveness is how and why successful fallacies remain undetected (that is less manifest or not manifest at all).

4.3 Why and How Fallacies Are Rhetorically Successful 4.3.1 Cognitive Underpinnings of Argument Evaluation From a cognitive perspective, and all else being equal, the evaluation of an argument can be modelled as a process in which two sets of information interact: on the one hand, the content of the speaker’s utterance, which contains the premise(s) and the conclusion, and on the other, the set of contextual information against which the

46

S. Oswald and T. Herman

speaker’s utterance is processed. In principle, this interaction can yield different outputs in the evaluation process, ranging from complete rejection of the claim to complete acceptance. This characterisation affords two different possible scenarios. In the first scenario, in which the set of contextual information that is mobilised to evaluate the claim and the arguments turns out to be somehow stronger (a notion we characterise more precisely below), the speaker’s standpoint is rejected. Conversely, in the second scenario, in which the set of contextual assumptions turns out to be weaker, the standpoint is likely to prevail and to enter the addressee’s cognitive environment.2 Let us further specify this picture, starting with a terminological discussion about the nature of information sets that interact in argumentative evaluation. In the evaluative stage of argumentation, where participants try to test the argumentation put forward by the speaker to figure out whether the conclusion should be accepted based on the available and supplied evidence,3 cognitive systems are processing these two information sets against one another: the content of argumentation on one hand and contextual information on the other. We propose to call the set of contextual information brought to bear in argumentative evaluation the critical context. The critical context contains information on, among other things, the speaker’s (possibly deceptive/manipulative/uncooperative) intentions, potential inconsistencies within the argumentation, potential discrepancies between the argumentation’s content and the addressee’s beliefs and values, etc. In short, the critical context is likely to contain any piece of relevant information that would help making the fallacious nature of the argumentation manifest. Let us now try to specify the criteria that are likely to influence argumentative evaluation, which, as previously mentioned, can result in either acceptance or rejection of the speaker’s standpoint. Earlier, we noted that acceptance denotes cases in which the content of the speaker’s argumentation is stronger than the critical context, and that rejection denotes cases in which the former is weaker than the latter. What exactly do we mean by ‘stronger’ and ‘weaker’? First, we have to note that these two comparative terms are to be understood as related to the selection of information: the stronger a piece of information, the more likely it is to be selected, processed, stored, etc., and, conversely, the weaker the piece of information, the less likely it is to be selected, processed, stored, etc. All else being equal, two main cognitive parameters are responsible for the selection of information in any cognitive process: the first has to do with its reliability, and the second with its processability. We define the first as epistemic strength and the second as accessibility or availability of critical information. Epistemic strength predicts acceptance of the conclusion in case the argument turns out to be epistemically superior to critical information 2 Sperber

and Wilson define the notion of cognitive environment as follows: “A cognitive environment of an individual is a set of facts that are manifest to him” (Sperber & Wilson, 1995, p. 39). We use this characterisation to denote quite loosely the set of information that an individual knows and is capable of bringing to consciousness. 3 We note, following Mercier and Sperber (Mercier & Sperber, 2009, 2011, 2017), that evaluation may be intuitive (that is, without consciously attending to the process) or reflective (that is, with reflective awareness of the evaluation).

4 Give the Standard Treatment of Fallacies a Chance …

47

sets: that is, the content of the argumentation put forward by the speaker is evaluated by the addressee as more likely to be true than whatever critical information he can come up with. We then define the second parameter (processability) as accessibility/availability of critical information and predict that failing to access critical information should increase persuasiveness.4 In lay terms, and to summarise, there are grounds to consider that an individual will end up accepting the conclusion of an argument when she finds its content particularly plausible in light of the criticism she can think of at that point. Another way of looking at this characterisation of rhetorical effectiveness is to envisage it as an issue of informational salience, in which persuasive strategies are characterised in terms of their strengthening and weakening possibilities. Strengthening strategies effectively foreground information that is compatible with the content of the claim/argument complex; this can be done by making such information epistemically strong and more accessible. Weakening strategies target critical information in order to background it; they achieve this by making it epistemically weaker and less accessible—if not altogether inaccessible. As argued elsewhere (e.g., Maillat & Oswald, 2009, 2011; Oswald, 2010, 2011, 2014, 2016) these parameters can be linked to Sperber and Wilson’s extent conditions of relevance (Sperber & Wilson, 1995, p. 125). While these were originally formulated in relevance theory to specify how addressees arrive at relevant interpretations of speaker meaning through inferential work, we propose that they are equally useful to any account of rhetorical effectiveness (see also Sect. 4.3.3). As per the principle of relevance [“Human cognition tends to be geared to the maximisation of relevance”, (Wilson & Sperber, 2002, p. 254)], these extent conditions can indeed be assumed to cover a larger ground and encompass virtually all cognitive activity, including the production and evaluation of argumentation. Under the reductionist perspective we explore here, the evaluation of argumentation is a process in which the cognitive system weighs the epistemic strength of a speaker’s argumentation against a (critical) subset of the addressee’s cognitive environment—this subset being itself likely to vary in terms of manifestness, depending on its accessibility. In such a framework, we therefore come to define rhetorical effectiveness as a function (i) of the epistemic strength of argumentation and (ii) of the accessibility/availability of critical information. Such a reductionist take on persuasion, we believe, is likely to open new avenues of research if we apply it to the study of fallacies and, more broadly, of argumentation (see Oswald, 2016).

4 Let

us note, in passing, that we assume that this happens on the condition that the content of the argumentation has some minimal degree of plausibility in the addressee’s cognitive environment. If the addressee outright rejects the content as being false, or if he simply does not understand it, then accessibility considerations in principle become irrelevant.

48

S. Oswald and T. Herman

4.3.2 Undetected Fallacies Are Arguments That Are Non-manifest Qua Fallacies Within this picture, and assuming that fallacies are normatively flawed arguments the rhetorical success of which rests on the non-manifestness of the flaw, we can hypothesise that fallacies operate specific cognitive constraints on the selection of information in the processes of evaluation. These constraints, in turn, are meant to focus the addressee’s attention on specific information sets and to prevent him from mobilising an optimal critical context during evaluation—or to prevent critical evaluation from starting at all. More specifically, we argue that fallacies operate a twofold cognitive constraint on information processing. On the one hand, they aim at strengthening ‘fallacy-friendly’ information sets, namely those sets of information that support or are compatible with the content of the fallacy. On the other, they aim at obfuscating critical information, or at least at weakening the chances that critical information is brought to bear at the evaluative stage. Saying that a fallacious argument has been rhetorically effective amounts to saying that either it has withstood critical evaluation or that the addressee has accepted it without any critical evaluation. In the first scenario, the result is that the content of the argumentation is either epistemically compatible with or epistemically stronger than the critical context against which it is tested. The second scenario can be interpreted in terms of the presence or absence of said critical context. If the fallacy prevails, it could simply be because no critical context is present at the time to defeat it or because the addressee finds no motivation to supply such a critical context; in this case what is at stake is the accessibility of the critical context. For theoretical and conceptual clarity, we should point out that fallacies are presumably no different from normatively sound arguments when we look at them from a cognitive perspective: both types of arguments are processed by general inferential mechanisms which regulate the selection of information following considerations of relevance, construed in terms of epistemic strength and accessibility. To take van Eemeren’s analogy with cats and dogs, fallacies and reasonable or sound arguments are not different animals: instead, fallacies are to be thought of as bad dogs while reasonable or sounds arguments are to be thought of as good dogs (van Eemeren, 2015). This analogy also makes sense from a cognitive perspective, as there are no solid reasons (empirical or otherwise) to consider that they rely on different cognitive processes. Furthermore, the unwitting addressee does not identify fallacies as such, meaning that their evaluation unfolds in the standard way—which is, admittedly, also part of the reason they are effective. However, they do differ in at least two respects, namely (i) their communicative status (in the case of deliberate fallacies, or fallacies by design) and (ii) their foregrounding/backgrounding dynamics. Many researchers have noted that fallacies can be treacherous, misleading and/or deceptive. Aristotle, in his Sophistical refutations, noted that fallacies many times take the appearance of valid deductions, and that this is at least partly why they can fool people (see Aristotle, 1955; Hansen, 2015). Bentham (1962) also highlighted the deceptive nature of fallacies, as they cause people

4 Give the Standard Treatment of Fallacies a Chance …

49

to have false beliefs. In contemporary research, Walton proposed to define a fallacy as “a deceptively bad argument that impedes the progress of a dialogue” (Walton, 1995, p. 256), thereby focusing his definition on both the deceptive and disruptive nature of fallacies. Finally, Tindale (2007) offers a view that is akin to a distinction made by Plantin (1995): fallacies can be deceptive, but they are not necessarily so, since they may arise as a consequence of errors in judgement or appreciation, which are accidental. To resolve the tension, Plantin proposes to distinguish sophisms from paralogisms on the criterion of intention: sophisms are intentional, and thus deliberately deceptive, while paralogisms are accidental, and thus not intended to deceive. All this research points to the fact that fallacies can be deceptive when speaker mean them as such: a corollary of the notion of fallacies by design is that they do not have communicative status—meaning the deceptive intention they try to fulfil must remain covert. Normatively acceptable arguments, conversely, are overt: they are formulated as sincere and acceptable arguments that are meant to convince, and they are precisely meant to be identified as sincere and acceptable arguments that are meant to convince. This is the first notable difference between fallacies and normatively sound arguments. The second difference, which we relate to the backgrounding/foregrounding dynamics of both types of arguments, has to do with the selectional constraints they operate on information sets. As noted, fallacies by design need to remain covert, which requires them to weaken the chances that a critical context will be mobilised or to make sure, it it is mobilised, that it is an epistemically weak one: we should thus expect them to try to keep critical information as weak as possible (in epistemic terms) and as inaccessible as possible (in terms of processing effort). In other words, deliberate fallacies try to sink the competition. This is not the case with normatively sound arguments: in fact, the best arguments, from a normative perspective, are precisely those that prevail after careful submission to doubt. From a dialectical perspective, furthermore, the best arguments are those that have withstood all potential objections—and this requires systematic and relentless mobilisation of critical information. So, in the end, normatively sound arguments encourage competition and promote critical discussion. This, we contend, translates into diverging selectional constraints on information at the cognitive level. In this framework, going beyond their normative shortcomings to assess their cognitive operability, we can therefore reinterpret known fallacies as specific strengthening and weakening strategies. Some will focus on strengthening contents in terms of epistemic strength, while others will devote their resources to weakening critical information in terms of its accessibility, and so on. For example, ad hominem fallacies reinterpreted in this fashion are weakening strategies meant to decrease the epistemic strength of a piece of information by discrediting its source. Mirroring this dynamic in an opposite direction, ad verecundiam fallacies try to implement strengthening strategies so as to increase the epistemic strength of fallacy-friendly information by giving credit to its source. Red herrings would also be weakening strategies, with the difference that what is weakened is the addressee’s likelihood of accessing critical information, as they effectively divert attention away from the argument’s problematic content, and thus away from the representation of relevant critical information.

50

S. Oswald and T. Herman

What could potentially lie ahead of us in this project, then, is a complete reassessment of existing work in fallacy research, which brings together the normative and the rhetorical side of argumentation. Before we discuss an example of how these constraints might play out in a context of persuasion, we need to add one layer to our rhetorical account of fallacious argumentation, which will take into account the interrelatedness of language, cognition and social context in argumentative contexts (see also Oswald & Herman, 2016). This will eventually lead us to formulate a clearer stance on what rhetorical effectiveness is in the framework we defend.

4.3.3 Pragma-Linguistic Constraints on Argumentation and Rhetorical Effectiveness The framework outlined above considers that cognitive mechanisms of argumentative evaluation may be disrupted because of the selectional constraints on information typically enforced through the use of fallacies. This, we contend, happens by the operation of strengthening and weakening strategies, which translate into the foregrounding and backgrounding of different sets of information, as just seen. While we so far hinted at the principled ways in which this might happen at the cognitive level, namely through the manipulation of epistemic strength and accessibility, we have not yet gone into the details of how exactly these strategies are realised, notably in terms of their verbal formulation. But looking at them first requires us to think about the different types of inferences that are performed in argumentative evaluation. At the core of our investigation lies the relationship between two types of inference: pragmatic inference (PI) and argumentative inference (AI). PI is an inference about meaning, which fulfils the interpretative goal of identifying speaker meaning, and which takes as input verbal and contextual information, in order to deliver an output in the form of a representation of speaker meaning. AI is about the acceptability of a justificatory link, which fulfils the evaluative goal of assessing the quality of argumentation, and which takes as input a representation of speaker meaning together with contextual information, in order to deliver an output in the form of an evaluative representation between a conclusion and reasons to accept it (see Oswald, 2018). Both PI and AI converge on one crucial dimension. Even if argumentation theory usually deals with questions related to argument quality and pragmatics is traditionally concerned with issues of meaning, both can indeed be taken to share one concern, namely curiosity for the mechanisms by which representations end up entertained by individuals through communication. Accordingly, in both accounts, quite minimally, the main issue is about explaining why and how a given representation becomes part of the individual’s cognitive environment. We have submitted earlier that the cognitive mechanisms invoked to account for PI are similar to those accounting for AI. It is now time to explore this idea more in detail, and we do so by first considering the relationship between PI and AI.

4 Give the Standard Treatment of Fallacies a Chance …

51

There are good reasons to consider that AI depends on PI. The first of them comes from cognitive considerations. From the above characterisations, it seems indeed that what AI partly takes as input, PI supplies as output—i.e. a representation of speaker meaning. Dan Sperber and his colleagues, in their work on epistemic vigilance, furthermore note that “comprehension of the content communicated by an utterance is a precondition for its acceptance” (Sperber et al., 2010, p. 367). This means that before one can assess whether a claim follows from supporting premises, one needs to be able to represent and understand the content of both the premises and the claim, since very little, if anything at all, is to be gained from an argumentative articulation of meaningless statements. The second line of inquiry providing evidence in favour of the idea that AI feeds on PI comes from rhetorical scholarship, which has long articulated a claim that rests on the very dependence of AI on PI: the way one frames an argument, in terms of the verbal choices one operates, can have significant impact on its chances of success. Rhetoric, classically construed as the study of the best available means of persuasion (Aristotle, 1995), has been in the business of studying the specificities of formal linguistic features of argumentative messages (syntactic structures, tropes, figures of speech, etc.) since its inception. In his work, Tindale motivates this idea and offers countless examples to show how different formulations might appeal to different audiences, thereby acknowledging the role PI has on AI (see Tindale, 1992, 2007, 2015). In some dialectically-inspired accounts, research on strategic manoeuvring, which incorporates a rhetorical component in the extended pragmadialectical model of argumentation (van Eemeren, 2010), points to this very idea in its discussion of the notion of presentational devices, which are linguistic choices made to enhance the persuasiveness of the message. Very recently, Schumann and her colleagues (Schumann, Zufferey, & Oswald, 2019) offered the first experimental study of the straw man fallacy to specifically cater for some linguistic and pragmatic criteria which influence its acceptability, i.e., its undetectability, showing decisive links between linguistic formulation and argumentative evaluation. All in all, there is overwhelming consensus in argumentation scholarship around the role of language and interpretation on persuasiveness, and thus indirectly on argument evaluation. In principle, hence, indicators converge to establish the relevance of accounts of utterance interpretation for the study of argumentation. The notion of pragmatic inference and its theorising owes a great deal to Grice’s pioneering pragmatic work on meaning (collected in his posthumous 1989 book), as his model of communication as a cooperatively rational conversational undertaking provided the building blocks of contemporary pragmatics. The starting point of his theory is the idea of semantic underdeterminacy, which follows from his (1957) distinction between natural and non-natural meaning. The idea of semantic underdeterminacy rests on the clear-cut distinction between what is said and what is meant. Grice provides a principled account (consisting of a cooperative principle and 4 conversational maxims which further specify what conversational cooperation should amount to) of how conversational participants are able to figure out that what speakers mean is many times different and more specific than what they say. Grice names this type of implicit meaning implicatures. Crucially, in his model, Grice considers that

52

S. Oswald and T. Herman

reaching a (naïve) interpretation of someone’s utterance involves inferring speaker meaning, which in turn boils down to recognising (and thereby fulfilling) a communicative intention. Implicatures are thus defined as contents which are implicitly speaker-meant and inferentially hearer-derived. Here we see that the construal of inference offered by Grice is not of an argumentative nature, as PI simply denotes the process by which implicit contents may be reached. Within relevance theory (Sperber & Wilson, 1995; Wilson & Sperber, 2012), PI is driven by considerations of relevance. That is, what makes addressees appropriately identify speaker meaning is the fact that their cognitive system identifies it as contextually relevant. In other words, understanding someone’s utterance amounts to identifying in which way that utterance is relevant in the context at hand. Relevance receives a technical definition, under this cognitive perspective, which characterises it as a relationship between processing effort (the use of cognitive resources to work out meaning) and cognitive effect (roughly speaking, the usefulness of an assumption, which relevance theory defines in epistemic terms). Relevant information, thus, is information that requires little processing efforts and yields significant cognitive effects. This effort-effect ratio is said to be responsible for the selection of information in any cognitive operation: once the system considers that expectations of relevance are fulfilled, it selects the output information and moves on to process other stimuli.5 An account such as the one introduced in relevance theory is instrumental to answering our research questions in at least two ways. First, the cognitive nature of the account allows us to characterise fairly precisely what we termed strengthening and weakening strategies: fallacies, in the end, manipulate the perceived relevance of information (relevance being defined as an optimal ratio between epistemic strength and processing effort). Second, using an account of PI to assess AI is an ideal move to assess the strong connection between the persuasiveness of people’s argumentations and their linguistic wording, which is of capital rhetorical importance. Earlier in this chapter (Sect. 4.2.2) we set out to explain how the standard treatment of fallacies could be salvaged by offering a rhetorical reinterpretation of its content. We are therefore now able to fully account for this reinterpretation within a rhetoric-pragmatic framework that caters for issues of persuasiveness. A final note, but an important one, is in order regarding what we mean by rhetorical effectiveness when we deal with fallacious arguments. The expression itself is vague, but intentionally so, as it may cover different kinds of disruptive argumentative behaviours. On the one hand, one could simply take rhetorical effectiveness to be a synonym of persuasiveness: rhetorically effective fallacies are those that manage to make sure that their conclusion ends up belonging to the addressee’s cognitive environment. This is the Aristotelian sense we mentioned earlier, whereby fallacies give the impression, and only the impression, of having successfully defended the claim. However, we want the expression to cover additional scenarios, such as distractions and failures to spot argumentative infelicities. If a speaker manages, for instance 5 We

refer the reader to additional work in relevance theory (e.g. Sperber & Wilson, 1995; Wilson & Sperber, 2002) for the detailed cognitive account of communicative processes.

4 Give the Standard Treatment of Fallacies a Chance …

53

through a fallacious ad hominem argument, to change the topic of the conversation to her own advantage, even if there is no actual persuasion about the standpoint defended by the ad hominem, we still want to call that rhetorically effective. More abstractly, we use the expression rhetorical effectiveness to also include scenarios in which a speaker manages to reach a conversational outcome where the fallacious moves she has made remain unidentified. This broader construal of rhetorical effectiveness is fully compatible with our pragmatically-driven account: explaining how information sets are selected and how fallacious arguments constrain this selection indeed affords such a broad construal because it is merely instrumental to explaining a range of different covert argumentative disruptions.

4.4 A Fallacious and Deceptive Erotetic Tour de Force The example we will discuss here comes from the 2009 film Public Enemies, directed by Michael Mann and adapted from Bryan Burrough’s non-fiction book Public Enemies: America’s Greatest Crime Wave and the Birth of the FBI, 1933–34 (2004).6 The book features the stories of many well-known gangsters, among others John Dillinger. In Chap. 9, Burrough recounts how Dillinger has been arrested and is held in Lake County Sheriff’s House and Jail, in Crown Point, Lake County, Indiana. Due to rumours about a possible escape, both the prosecutor, Robert Estill, and the female director of that prison, Sheriff Holley, request that Dillinger be transferred to a higher-profile prison, Indiana State Prison, in Michigan City. Dillinger does not want this, as an escape plan is more likely to succeed where he is, in Lake County, than where they want to send him, which is why he instructs his lawyer Louis Piquett to try to talk the judge out of granting the transfer request. We reproduce below the excerpt from the film, whose dialogical format is more practical to analyse:7 Context:

Dillinger is shackled to his chair. The courtroom is packed. Walls are lined with deputies holding rifles. Reporters scribble. The gavel quiets the crowd Piquett (on his feet) Your honor! Are we to have a hearing in accord with the laws of this nation, or is the State to be permitted to incite an atmosphere of prejudice? The very air reeks with the bloody rancor of malice. The clanging of shackles brings to our minds the dungeons the Czars, not the flag-bedecked of an American courtroom. I request the court to direct that those be removed forthwith! 6 According

to the author, “The conversations and dialogue in this book are taken verbatim from FBI reports, the Karpis transcripts, contemporary news articles, and the memories of participants.” (Burrough, 2004, preface). Following this disclaimer, we presume that the dialogue under analysis here has taken place in real life, even if beyond the author’s declaration, we have no guaranteed way of ascertaining that it has indeed been reproduced verbatim. 7 Script available at https://www.imsdb.com/scripts/Public-Enemies.html, last accessed 27.03.2019.

54

Robert Estill: Deputy Holley Piquett Judge Murray Robert Estill

Judge Murray Sheriff Holley Piquett Sheriff Holley Piquett

Sheriff Holley Judge Murray

S. Oswald and T. Herman

This is a very dangerous man, your Honor. And I’m responsible for the safeguarding the prisoner. Who are you?! Are you a lawyer? What right have you to address this court? Alright, remove the handcuffs from the prisoner. Your honor, we’d like to relocate the prisoner. Only Indiana State Prison in Michigan city can guarantee Dillinger will not escape. Sheriff Holley? I concur, your honor. Sheriff Holley, I think it’s a very nice jail you have right here. What makes you think there’s anything wrong with it? There isn’t anything wrong with my jail! It’s the strongest jail in Indiana. That’s what I thought. But of course, I don’t want to embarrass Mrs. Holley. I appreciate that she’s a woman and if she’s afraid of an escape… I’m not afraid of an escape. I can take care of John Dillinger or any other prisoner. Okay, Dillinger will stay here.

Piquett’s strategy to get Dillinger to stay in Lake County Jail is rhetorically interesting on a number of counts. First, it instantiates three crucial questions which are meant to take control of the discussion. Second, it instantiates at least two known and possibly three fallacious arguments. Third, it manages to get Piquett’s adversaries to do all the argumentative work—so to speak, Sheriff Holley walks into a trap on her own, and she is the one to provide the decisive piece of information that tips the scales in favour of Piquett and Dillinger. Our analysis begins with an assessment of the particular linguistic construction and unfolding of this strategy, in the hope of highlighting how and why the fallacies remain invisible to their addressees until it is too late. For this, we will look at the specific constraints that Piquett’s series of questions impose on the interpretation of his words.

4.4.1 A Double Rhetorical Trick Question Realising a False Dilemma We start with Piquett’s opening disjunctive question, which we reproduce in (1) below: (1) (a) “Are we to have a hearing in accord with the laws of this nation, or (b) is the State to be permitted to incite an atmosphere of prejudice?” We take (1) to be an instantiation of the fallacy known as false dilemma (see Boone, 1999; Tomi´c, 2013), as it indeed seems to us that the terms of the proposed alternative

4 Give the Standard Treatment of Fallacies a Chance …

55

are neither mutually exclusive nor exhaustive. From a normative perspective, therefore, this move is fallacious. Even if from a formal perspective (1) could be evaluated as deductively valid, the argument fails in that the disjunctive premise underlying the questions is false. We are thus not legitimated in drawing the first disjunct from the disjunction. There is, however, much more to this strategy if we think about how it is likely to be processed. (1) is uttered at a particular moment and in a particular context: the film stages Dillinger’s entrance in the courtroom first through sound (we hear shackles clinging) and then by a vertical pan shot spanning from the feet to the head of the gangster. These are actually the visual and auditory cues that prompt Piquett’s first question (1a). Thus, auditory, visual and verbal cues in this case converge to create a context in which Dillinger is represented as being the victim of unfair, borderline unlawful, treatment. Let us incidentally note that the order in which the two disjuncts of the false dilemma are uttered is in itself interesting. While in principle, and all else being equal, a premise such as “either A or B” could be reversed as “B or A” with no notable difference, we submit that this is unlikely to be the case in this context because the order of the clauses seems rhetorically significant. The question contained in (1a) is a rhetorical question whose answer should be ruled out as irrelevant due to its oddness, if not its absurdity.8 Yet, we surmise that it is likely to trigger interesting cognitive effects and thus to achieve relevance in a context like the one we just described. The situation itself (a shackled Dillinger, victim of a form of institutional indignity) actually licenses and legitimates the question. While (1b) could be considered as an attack against the institution, (1a) summons what should be standard and normal legal procedure (obeying the laws of the Nation) to confront what Piquett sees as an abnormal situation which does not seem to be compatible with the rule of law. The move orchestrated in (1a) seems to rest on a modus tollens that we reconstruct as follows: (2) (a) If the court acted in accord of the laws of this Nation, Dillinger would not be shackled. (b) Dillinger is shackled. (c) Therefore the court did not act in accord of the laws of this Nation. Since the conclusion is untenable for an institution whose mission is precisely to make sure that the laws of the Nation are observed, it becomes impossible to keep the prisoner shackled or to plead an exception to the application of the law—and the judge indeed complies with Piquett’s request. Regarding the linguistic formulation of this strategy, two important observations are in order. First, if Piquett had opened with (1b), he would have automatically assumed an antagonistic stance, which would have prompted the judge to either defend his court or to sanction the lawyer for the affront. In such a scenario, Piquett would have first supplied his own reading of the situation, and that would have opened his flank to criticism: volunteering one’s opinion inevitably comes with the risk of being criticised and refuted. But by expressing his opinion in (1b) instead of 8 The

answer to this question is obviously and unquestionably a very trivial ‘yes’.

56

S. Oswald and T. Herman

in (1a), Piquett presents it as a mere description of facts that seems incompatible with his recalling of the court’s mission (i.e., the rule of law referred to in (1a)). From a cognitive perspective, this amounts to weakening the likelihood that the judge will engage in criticism and/or sanctioning. Calling attention first to a state of affairs that he deems abnormal under the rule of law makes available a rhetorically interesting option: Piquett is now able to use this as evidence for the relevance of (1b)—and he needs (1b), as we show next. We venture that our second observation, which concerns the perceived relevance of (1a), makes this line of analysis more plausible. The acceptability of the conclusion (2c) on which (1a) rests is so ethically problematic that Piquett would not have even needed to formulate an alternative. He could have limited himself to requesting the court to remove the shackles based on (1a) only. Yet, (1b) allows him to strengthen the fundamental assumption that the context is unfair to Dillinger. Since the alternative is manifestly absurd (as (1b) is clearly to be rejected), the relevant point is not to know whether the State is allowed to incite an atmosphere of prejudice, but instead to make more accessible, and possibly epistemically stronger, the assumption that seeing Dillinger in shackles is actual evidence of said prejudice. In the immediate aftermath, as he grants the request, it is striking to see that the judge does not invalidate this interpretation: on the contrary, he confirms its relevance, thereby strengthening it. This is why (1b) becomes so difficult to refute. To summarise: through a double rhetorical trick question, Piquett manages to create a context in which Dillinger is relevantly portrayed as a victim of unfair treatment by a prejudiced State. Additionally, since the latter has manifestly incited prejudice, it becomes difficult for the judge to credibly refute (1b). Piquett has therefore managed to strengthen both the confrontational setting and Dillinger’s role of victim (additionally confirmed by Piquett’s following utterance “the very air reeks with the bloody rancor of intolerant malice”). This is fundamental in the lawyer’s overall rhetorical strategy, because it sets the stage for what comes next: for this we have to keep in mind that whatever follows will be processed against a context in which the court’s unfairness and Dillinger’s victimhood are highly salient assumptions.

4.4.2 More Rhetorical Questions to Strengthen the Context The deputy’s contribution is quickly attacked by Piquett through, again, two rhetorical questions: (3) “Are you a lawyer?” (invited answer: ‘obviously not’) (4) “What right do you have to address this court?” (invited answer: ‘obviously none’) The information-seeking appearance of questions (3) and (4) is manifestly irrelevant, since the answers are obvious. This means that their relevance should be found elsewhere. One option would be to consider that Piquett is out for an ad hominem

4 Give the Standard Treatment of Fallacies a Chance …

57

attack meant to disqualify the speaker as a credible—if authorised—speaker in the courtroom, but we think that the point of this attack actually lies elsewhere. In terms of their content, these two rhetorical questions are meant to prevent the deputy from addressing the court based on legal considerations (only legally sanctioned individuals may address the court). Thus, Piquett appears, again, through (3) and (4), to make sure that the court observes the laws of the Nation and that Dillinger receives a fair trial. Consequently, (3) and (4) are likely to be perceived as consistent with (1a) and (1b) because they share the same intent. Just as he fought the illegal shackling of Dillinger, he fights to prevent unauthorised individuals from speaking in court. Two different strategies are at play here, one directed at his own ethos, the other at the image of the court he wants to convey: while he poses as the defender of the law, the court seems to have contempt for it, in light of its behaviour, which Piquett is very quick to highlight. In cognitive terms, Pickett is, once again, trying to make sure that whatever happens next is assessed against a context that has strengthened, or foregrounded, the biased and prejudiced nature of the court. This second instantiation of the strategy is cognitively advantageous, simply because it is “more of the same”: yet again, Piquett shows that the court is so hostile towards Dillinger that it allows attendants (in this case, the deputy) to engage in illegal behaviour. And this, as per our cognitive account, is a cognitively optimal move, since it proves easy to process, making that particular set of contextual information highly accessible.

4.4.3 The Straw Man Question The previous sets of questions were preparatory moves for the final touch of Piquett’s erotetic strategy, which is ultimately about manipulating his opponents into admitting that Lake County Jail is a very safe prison. If he manages to do that, then he ipso facto refutes the grounds to transfer Dillinger to a more secure prison. Moreover, it would be rhetorically more powerful that this piece of information emerges as an admission from his opponents, rather than as one of his own arguments or refutations. Indeed, if he succeeds, not only will the opponents provide a refutation to their own standpoint; they will also, in the process, look bad for not having worked out their request in a sound way. And for that, he crucially needs to obfuscate the fact that the admission that Lake County Jail is very secure functions as a strong counter-argument to their request for transfer. We will see that he secures that outcome by managing to deflect the function and relevance of that particular admission towards a defence following a character attack on Sheriff Holley. In particular, we will show how Piquett does this through a straw man fallacy, which is a fallacy of refutation in which an arguer’s position is misrepresented in order to be attacked more easily (see e.g., Aikin & Casey, 2011; Lewi´nski & Oswald, 2013; Oswald & Lewi´nski, 2014). In the dialogue, Robert Estill, the prosecutor, states the following:

58

S. Oswald and T. Herman

(5) “Only Indiana State Prison in Michigan city can guarantee Dillinger will not escape.” Sheriff Holley, when asked, explicitly concurs, and this is followed by another question by Piquett (6): (6) “What makes you think there’s anything wrong with it [Lake County Jail]?” Through this question, Piquett misrepresents his interlocutors’ position: they have not conceded that there is something wrong with Lake County Jail, but that Indiana State Prison is better, which are two different things. This is a blatant misrepresentation, if only because, topic-wise, Piquett’s question is about Lake County Jail, while the prosecutor’s assertion is about Indiana State Prison. However, we believe—and this is indeed what happens—that Piquett’s straw man has prospects of being rhetorically effective, that is, of disrupting the conversation to achieve the intended result of securing an admission. We can support this analysis through the following five observations. First, as defended in what precedes, Piquett utters (6) in an immediate context of suspicion of bias against Dillinger. The lawyer has managed to show that the court is prejudiced against the prisoner, which makes the possibility of injustice extremely salient: in such a context, perhaps the prosecution’s request for a prison transfer can even be seen as an unjustified and useless manoeuver. Second, (6) is a loaded question which presupposes two important things: (7) There is something wrong with Lake County Jail (8) The director of Lake County Jail believes that there is something wrong with her prison These presuppositions are relevant on two counts: first, they echo the hypothesis we just evoked, namely the possibility of injustice and the fact that there might be something fishy behind the transfer request. Nevertheless, we think the rhetorical effect of this presupposition is actually to be found at the conversational level: for obvious face reasons, Sheriff Holley cannot come across as someone who believes that her prison is unsafe, and thus the presupposition in (8) strengthens the need for a defence on her behalf.9 She in fact takes on the presuppositional challenge, unfortunately for her without realising that she is actually falling into the trap that will fulfil Piquett’s ulterior motive (to get Dillinger to stay at Lake County Jail). By reacting to (8) through a refutation, she provides Piquett with the admission he was looking for from the beginning. Third, the strength of the straw man, together with the likelihood that it will not be detected—which, in our view, rests on the interpretation of the straw man as an attack that needs to countered by Sheriff Holley, and not as a call for an admission—are increased by the attributive format of the question. Piquett uses the second person 9 The

reference to her gender—and, implicitly, to the associated stereotype of inferiority—could function as an additional constraint designed to increase the chances that Sheriff Holley takes on the presupposition and responds to it in a way that makes her admit that her prison is safe. This is all Piquett needs to conclude that there are no grounds whatsoever to transfer Dillinger.

4 Give the Standard Treatment of Fallacies a Chance …

59

pronoun “you” to address the Sheriff directly. In so doing, he shifts the burden of proof to Sheriff Holley, who finds herself in the position of having to disprove a misattribution for which she is not even responsible. Fourth, we observe a strange inversion in the dialectical roles due to the misrepresentation Piquett operates in the straw man: at the end of the exchange, the lawyer is in the unnatural position of defending Lake County Jail and in fact appears to be more defensive of this prison than its own director. And this, rhetorically speaking, seems extremely effective, because it casts an unfavourable light on the Sheriff. Fifth, there is some semantic proximity between what Sheriff Holley says and Piquett’s misrepresentation. The reason for this is that thinking that something is wrong with Lake County Jail could be a possible entailment of considering that only Indiana State Prison can guarantee Dillinger will not escape. The problem, of course, is that this is arguably not a necessary entailment, although the use of the adverb ‘only’ suggests an exclusive reading favouring the entailment. This goes to show that Piquett’s move is not completely off-mark, which is also why it has prospects of being rhetorically effective. Piquett’s overall argumentative strategy is twofold: on the one hand, his erotetic scheme heavily relies on the audience’s inferential processes to work out the argumentative thread. All questions are either leading, rhetorical or treacherous, and all of them are designed to be rhetorically favourable to Piquett and Dillinger. Some of them target the construction of a context in which he and his client are victims of the court’s unlawful treatment, and others are meant to absolve Piquett from the need to state his position and to formulate complete arguments. On the other hand, these questions are meant to put Piquett’s interlocutors in a defensive position, since in all three cases Piquett attributes a standpoint to his interlocutor, what is more a standpoint that the interlocutors need to exonerate themselves from. The judge has to prevent the crowd from thinking that he is presiding an unlawful court, the deputy needs to disassociate from Piquett’s implication that he is failing to observe the law by addressing the court, and Sheriff Holley needs to fight the misattribution that she believes her prison to be unsafe. Thus, instead of needing to argue, Piquett makes the burden of proof rest on his interlocutors’ shoulders. This strategy culminates with the admission that Lake County Jail is a safe prison, which is all he needed to make his case. This is a complex erotetic strategy, but we do think that it lends itself to the kind of framework we developed earlier. In cognitive terms, Piquett leads the show because he is able to constrain which pieces of information become relevant in his interlocutor’s minds, as they answer his questions. In (1a) and (1b), he generates expectations of relevance for (i.e., he tries to increase the subsequent accessibility of) any content related to the confrontational and prejudiced attitude of the court and to Dillinger’s assumed victimhood. The context that is then available prominently contains these assumptions. In (3) and (4), he consolidates this context, thereby securing the presence of the same discriminatory assumptions in the cognitive environments of the members of the audience. Finally, in (6), which presupposes (7) and (8), he makes sure that Sheriff Holley interprets the question as an attack, and not as a way of eliciting the admission that Lake County Jail is a very secure prison. He does this by

60

S. Oswald and T. Herman

making the attack interpretation extremely accessible (in light of the first two sets of questions and of the presuppositions). The need for defensive action is thus overwhelming, which is how she ends up unwittingly admitting to what Piquett wanted her to admit. The rhetorical effectiveness of this erotetic strategy is multifaceted. On the one hand, Piquett has managed to convince everyone that Lake County Jail is safe by getting its director to admit that there is nothing wrong with it. In so doing, he has managed to establish that there is no need to transfer Dillinger. On the other hand, he has successfully managed to conceal the means by which he wanted to achieve that, by distracting the audience into thinking that the dialogue was about a personal attack, when in fact it was about getting Sheriff Holley to volunteer a decisive piece of information. This goes to show that the complexity of the notion of rhetorical effectiveness, as a theoretical construct, stands to gain from the type of account presented here, which captures different subtypes of rhetorical effectiveness, thereby achieving descriptive and explanatory adequacy.

4.5 Conclusion In this chapter, we have tried to rehabilitate the definition of fallacy given in the standard treatment by showing that what is a problem for the normative question (i.e., its psychological component) is actually an advantage for the explanatory question related to rhetorical effectiveness. We have shown that the cognitive evaluation of arguments can be represented as boiling down to figuring out how compatible sets of information (here the content of the argumentation and the critical context) are, so as to determine whether the conclusion of the argumentation should be included in the cognitive environment of the addressee. We have further established that said compatibility is determined along two cognitive parameters: epistemic strength and accessibility of critical information. Through our pragma-rhetorical analysis of a complex erotetic argumentative strategy, we have tried to show that our cognitive approach to fallacies has interesting prospects of contributing to argumentation scholarship in novel ways. Our discussion yielded results which we think are indeed methodologically valuable, on at least three different levels. First, from a descriptive vantage point, our account allows us not only to study the specificities of each known fallacy, but also to observe how these specificities might be realised through different linguistic choices. Second, from an explanatory perspective, we are now able to supply a principled—and furthermore testable—account of why these linguistic choices are likely (or unlikely), in a particular context, to be rhetorically successful. Third, we can now start thinking about different types of rhetorical effectiveness, to be documented within one and the same theoretical framework. In terms of the development of the discipline, this pleads for the need for further integrative inquiries developed along the lines introduced here.

4 Give the Standard Treatment of Fallacies a Chance …

61

References Aikin, S. F., & Casey, J. (2011). Straw men, weak men, and hollow men. Argumentation, 25(1), 87–105. Aristotle. (1955). On sophistical refutations; On coming-to-be and passing away; On the cosmos (Vol. 3; E. S. Forster & D. J. Furley, Trans.). Cambridge, MA: Harvard University Press. Aristotle. (1995). The complete works of Aristotle: The revised Oxford translation. Princeton, NJ: Princeton University Press. Bentham, J. (1962). The handbook of political fallacies (Vol. 69). New York, NY: Harper & Brothers. Boone, D. N. (1999). The cogent reasoning model of informal fallacies. Informal Logic, 19(1), 1–19. Burrough, B. (2004). Public enemies: America’s greatest crime wave and the birth of the FBI, 1933-34. London: Penguin Press. Correia, V. (2011). Biases and fallacies: The role of motivated irrationality in fallacious reasoning. Cogency: Journal of Reasoning and Argumentation, 3(1), 107–126. Correia, V. (2014). Biased argumentation and critical thinking. In T. Herman & S. Oswald (Eds.), Rhétorique et cognition. Perspective théoriques et stratégies persuasives. Rhetoric and Cognition. Theoretical perspectives and persuasive strategies (pp. 89–110). Bern: Peter Lang. Grice, H. P. (1957). Meaning. Philosophical Review, 66(3), 377–388. Hamblin, C. (1970). Fallacies. London: Methuen. Hansen, H. (2015). Fallacies. In E. N. Zalta (Ed.), The Stanford encyclopedia of philosophy (Summer 2015). Retrieved from https://plato.stanford.edu/archives/sum2015/entries/fallacies/. Jackson, S. (1996). Fallacies and heuristics. In J. van Benthem, F. van Eemeren, R. Grootendorst, & F. Veltman (Eds.), Logic and argumentation (pp. 101–114). Amsterdam: Royal Netherlands Academy of Arts and Sciences. Johnson, R. H. (2014). The rise of informal logic: Essays on argumentation, critical thinking, reasoning and politics (Vol. 2). Windsor, ON: University of Windsor. Johnson, R., & Blair, J. A. (2006). Logical self-defense. New York, NY: International Debate Education Association. Lewi´nski, M., & Oswald, S. (2013). When and how do we deal with straw men? A normative and cognitive pragmatic account. Journal of Pragmatics, 59, Part B, 164–177. https://doi.org/10. 1016/j.pragma.2013.05.001. Maillat, D., & Oswald, S. (2009). Defining manipulative discourse: The pragmatics of cognitive illusions. International Review of Pragmatics, 1(2), 348–370. Maillat, D., & Oswald, S. (2011). Constraining context. A pragmatic account of cognitive manipulation. In C. Hart (Ed.), Critical discourse studies in context and cognition (Vol. 43, pp. 65–80). Amsterdam: John Benjamins. Mercier, H., & Sperber, D. (2009). Intuitive and reflective inferences. In J. Evans & K. Frankish (Eds.), In two minds: Dual processes and beyond (pp. 149–170). Oxford: Oxford University Press. Mercier, H., & Sperber, D. (2011). Why do humans reason? Arguments for an argumentative theory. Behavioral and Brain Sciences, 34(02), 57–74. https://doi.org/10.1017/S0140525X10000968. Mercier, H., & Sperber, D. (2017). The enigma of reason. Cambridge, MA: Harvard University Press. Oswald, S. (2010). Pragmatics of Uncooperative and Manipulative Communication. Université de Neuchâtel. Oswald, S. (2011). From interpretation to consent: Arguments, beliefs and meaning. Discourse Studies, 13(6), 806–814. https://doi.org/10.1177/1461445611421360e. Oswald, S. (2014). It is easy to miss something you are not looking for: A pragmatic account of covert communicative influence for (critical) discourse analysis. In C. Hart & P. Cap (Eds.), Contemporary studies in critical discourse analysis (pp. 97–119). London: Bloomsbury Publishing.

62

S. Oswald and T. Herman

Oswald, S. (2016). Rhetoric and cognition: Pragmatic constraints on argument processing. In M. Padilla Cruz (Ed.), Relevance theory: Recent developments, current challenges and future directions (Vol. 268, pp. 261–285). Amsterdam: John Benjamins. Oswald, S. (2018). Pragmatic inference and argumentative inference. In S. Oswald & D. Maillat (Eds.), Argumentation and Inference: Proceedings of the 2nd European Conference on Argumentation, Fribourg 2017 (Vol. 2, pp. 615–629). London: College Publications. Oswald, S., & Herman, T. (2016). Argumentation, conspiracy and the moon: a rhetorical-pragmatic analysis. In M. Danesi & S. Greco (Eds.), Case studies in discourse analysis (pp. 295–330). Munich: Lincom Europa. Oswald, S., & Lewi´nski, M. (2014). Pragmatics, cognitive heuristics and the straw man fallacy. In T. Herman & S. Oswald (Eds.), Rhétorique et cognition: Perspectives théoriques et stratégies persuasives—Rhetoric and cognition: Theoretical perspectives and persuasive strategies (pp. 313–343). Bern: Peter Lang. Plantin, C. (1995). L’argument du paralogisme. Hermès, 1, 245–262. Schumann, J., Zufferey, S., & Oswald, S. (2019). What makes a straw man acceptable? Three experiments assessing linguistic factors. Journal of Pragmatics, 141, 1–15. Sperber, D., Clément, F., Heintz, C., Mascaro, O., Mercier, H., Origgi, G., et al. (2010). Epistemic vigilance. Mind and Language, 25(4), 359–393. https://doi.org/10.1111/j.1468-0017.2010. 01394.x. Sperber, D., & Wilson, D. (1995). Relevance: Communication and cognition. Oxford: Blackwell. Tindale, C. W. (1992). Audiences, relevance, and cognitive environments. Argumentation, 6(2), 177–188. https://doi.org/10.1007/BF00154324. Tindale, C. W. (2007). Fallacies and argument appraisal. Cambridge, MA: Cambridge University Press. Tindale, C. W. (2015). The philosophy of argument and audience reception. New York, NY: Cambridge University Press. Tomi´c, T. (2013). False dilemma: A systematic exposition. Argumentation, 27(4), 347–368. van Eemeren, F. H. (2001). Fallacies. In F. H. van Eemeren (Ed.), Crucial concepts in argumentation theory (pp. 135–164). Amsterdam: Amsterdam University Press. van Eemeren, F. H. (2010). Strategic maneuvering in argumentative discourse: Extending the pragma-dialectical theory of argumentation. Amsterdam: John Benjamins. van Eemeren, F. H. (2015). Reasonableness and effectiveness in argumentative discourse: Fifty contributions to the development of pragma-dialectics. Amsterdam: Springer. van Eemeren, F. H., Garssen, B., Krabbe, E. C. W., Henkemans, A. F. S., Verheij, B., & Wagemans, J. H. M. (2014). Handbook of argumentation theory (2014th ed.). New York, NY: Springer. Walton, D. (1995). A pragmatic theory of fallacy. Tuscaloosa: University of Alabama Press. Walton, D. (2010). Why fallacies appear to be better arguments than they are. Informal Logic, 30(2), 159–184. Willard, C. (1995). Failures of relevance. In H. V. Hansen & R. C. Pinto (Eds.), Fallacies. Classical and contemporary readings (pp. 145–158). Pennsylvania: The Pennsylvania State University Press. Wilson, D., & Sperber, D. (2002). Relevance theory. UCL Working Papers in Linguistics, 13, 249–287. Wilson, D., & Sperber, D. (2012). Meaning and relevance. Cambridge; New York: Cambridge University Press.

Chapter 5

Argument Evaluation in Philosophy: Fallacies as Strategic Maneuvering Federico E. López

5.1 Introduction The future historian of “thought and expression” in the twentieth century will no doubt record with some amusement the ingenious trick, which some of the philosophical controversialists of the first quarter of our century had, of labelling their opponents’ views ‘fallacies’. He may even list some of these alleged fallacies for a certain sonority which their inventors embodied in their titles: the fallacy of initial predication, the fallacy of simple location, the fallacy of misplaced concreteness, the naturalistic fallacy. (Frankena, 1939)

As it is usually recognized, one of the philosophical motives behind the recovery of argumentation theory in the twentieth century was the attempt to renew and broaden the very concept of reason or rationality. In fact, Chaim Perelman and Lucie Olbrechts-Tyteca begin their influential and founding book by claiming that the publication of a treatise on argumentation—which takes into account the Greek tradition of rhetoric and dialectics—“constitutes a break with a concept of reason and reasoning due to Descartes which has set its mark on Western philosophy for the last three centuries.” (1969, p. 1). This new conception was meant to be a result of a way to proceed, similar to that used by formal logicians. Gottlob Frege, as Perleman reminds us (1979, p. 9), analyzed the reasoning used by mathematicians and, in so doing, he succeeded in making an important progress in his discipline. According to Perelman, it was then time to analyze the reasoning used in philosophy and other humanities, and to reconstruct the underlying rationality there involved, a rationality that could not be reduced to a formal proof nor to a scientific experiment. In that way, a double relation between philosophy and argumentation theory was suggested. In Perelman words, “If philosophy makes it possible to clarify and render precise the F. E. López (B) Centro de Investigaciones Filosóficas, Facultad de Humanidades y Ciencias de la Educación, Universidad Nacional de La Plata, Ensenada, Buenos Aires, Argentina e-mail: [email protected] © Springer Nature Switzerland AG 2020 F. H. van Eemeren and B. Garssen (eds.), From Argument Schemes to Argumentative Relations in the Wild, Argumentation Library 35, https://doi.org/10.1007/978-3-030-28367-4_5

63

64

F. E. López

basic notions of rhetoric and dialectic, the rhetorical perspective makes it possible to understand the philosophical enterprise itself better (…).” (1979, p. 50). In this vein, this paper attempts to contribute to such understanding by adopting a pragmadialectical perspective which, following van Eemeren (2010), assumes also a rhetorical dimension. In that way, it attempts to reflect on some peculiarities of philosophical controversies as critical discussions. In particular, this paper focuses on one of the evaluative strategies that philosophers use in their exchanges, namely, the charge of committing a fallacy. The study of fallacies has been an important topic in the history of logic and argumentation theory and the use of the several kinds of fallacies is a common practice in different contexts. However, it is in the context of philosophical debates that they appear more frequently, although usually in a controverted way. This controversial use of fallacies is brought to the fore in Frankena’s words which focus on the use of fallacies within philosophy. In describing such use as an ingenious trick, he draws attention to the fact that, as philosophical evaluative strategies, fallacies can be effective but misguided moves: if those alleged fallacies are indeed inventions one of the controversialists came up with, it may as well be that the charge of committing one, made by one participant against the other, is itself fallacious. In order to gain a better understanding of the use of fallacies in philosophical critical discussions—and to illustrate some of its problems—this article addresses the case of the genetic fallacy. This fallacy, which allegedly consist in confusing the genesis of an idea with its validity, repeatedly appears in the context of analytical philosophy and was particularly important since the firsts decades of twentieth century. Along with other fallacies such as the naturalistic fallacy, the genetic fallacy is commonly used against some stances and approaches in philosophy and specially in ethics and has generated different replies. Indeed, around the middle of the twentieth century, some controversy on the genetic fallacy aroused. The analysis of this controversy, which is carried out in this paper, shows some of the peculiarities of the use of fallacies in philosophy and some of his problems. With the aim of better understanding it, the charge of committing a (genetic) fallacy will be analyzed as strategic maneuvering (van Eemeren, 2010). Such analysis will make possible to appreciate the dialectical and rhetorical functions that that move fulfils. Accordingly, and to begin with, the problem of argument evaluation in philosophy is presented in a general way. Then, a characterization of this fallacy is provided, noting some difficulties in distinguishing between correct and fallacious uses of what can be considered genetic argumentation. In the second place, some historical evidence is examined by revisiting the controversy given rise to over the genetic fallacy around the nineteen fifties and sixties. This analysis suggests that the charge of committing a genetic fallacy has been frequently considered as strategic maneuvering. Consequently, an account of charging the other party of committing a genetic fallacy as strategic maneuvering is presented and some ways in which such maneuver could derail are explicitly stated. To conclude, some general remarks will be made regarding the use of (certain) fallacies in philosophical controversies and the value of the concept of strategic maneuvering in such context.

5 Argument Evaluation in Philosophy: Fallacies …

65

5.2 Argumentation in Philosophy: The Problems of Evaluation Standars As it is usually recognized, the relation between philosophy and argumentation is as ancient as philosophy itself. Indeed, even though influential philosophers as Friedrich Nietzsche, Walter Benjamin, or even Ludwig Wittgenstein,1 have sometimes used non argumentative tools to express their ideas and stances, a large part of philosophers are inclined to see essential bonds between philosophy and argumentation. Even the concept of argument in the narrow sense of premise-conclusion compound, has been understood as the crucial and only methodological tool that philosophers can fall back on, inasmuch as the argument would be a sort of methodological equivalent to the scientific experiment: whatever point of view philosophers long for maintaining, they should prove it or show it by means of a sound argument. This analogy between philosophical arguments and scientific experiments could be taken further. In the context of nineteenth and twentieth centuries philosophy of science, the concept of crucial experiment was usual and, unsurprisingly, polemic. Leaving aside the problems that defining such a concept involves (see Lakatos, 1974), a crucial experiment was told to be an experiment that plays a major part “in deciding between theories and establishing a basis for further research” (Roll-Hansen, 1989, 304). In the history of philosophy there are arguments that were considered as crucial arguments in a resembling sense, that is to say, arguments which, at least for some time, were considered as well established and that consequently were taken as starting points not only for a particular philosophical dialogue but also for the entire field. Kant’s objection against Anselm’s ontological argument, Sellars’s critique of “the Myth of the given”, or Quine’s rejection of the “two dogmas of empiricism,” could be great examples of argumentations that are usually considered as crucial. In these cases, arguments that are used to refute an opponent’s point of view have been considered also as a bedrock for further research. To mention another example, the use of the very idea of naturalistic fallacy has played a similar role in promoting some kinds of inquiries within philosophical ethics and in discouraging others. However, things are not that simple. There is a trait of philosophical discussions that makes them problematic. The status of the criteria used to evaluate an argument in philosophy and, for instance, to determine if an argument counts as a crucial argument, are usually disused and decided within the context of certain philosophical theory or tradition. For instance, the so-called thought experiments, which have a long history and are frequent in contemporary analytical philosophy, are sometimes questioned in its probative force. Something similar could be said about Kantian transcendental arguments, which have been refused by philosophers belonging to other traditions, as Richard Rorty. In a certain sense, it could be said that Philosophy looks like a pre-paradigmatic discipline, to use Kuhn’s term, in which there is no 1 In 1912, speaking about Wittgenstein, Russell writes: “I told him he ought not simply to state what

he thinks true, but to give arguments for it, but he said arguments spoil its beauty, and that he would feel as if he was dirtying a flower with muddy hands. … I told him I hadn’t the heart to say anything against that, and that he had better acquire a slave to state the arguments” (Monk, 1996, p. 264).

66

F. E. López

agreement about which the acceptable methods and assumptions are. Analytical Philosophy could be an exception to the extent that it was considered as “philosophy conducted within a paradigm” (Levi, 2003, p. 284). However, the problem returns as soon as philosophers engage in a dialogue between traditions. This fact could be interpreted as showing that there is no way to decide if a philosophical argument is a good argument or, in the best-case scenario, as implying that only internal evaluation is possible. In this vein, Perelman remind us that for some philosophers, “what is specific to philosophical controversy is that it is entirely on ad hominem argumentation, since it can only be based on the theses explicitly recognized by the philosopher himself, which he cannot deny without contradicting himself.” (1979, p. 50). In this context, the concept of fallacy could be useful. As it is usually understood, a fallacy is a violation of some kind of logical rule, and consequently, the detection of fallacies could be a non-controverted practice. However, the use of fallacies has its own problems which will be considered in the following sections.

5.3 The Controversy About the Genetic Fallacy 5.3.1 What Is the Genetic Fallacy? As mentioned before, the detection of fallacies is a common practice among philosophers. One of the fallacies that is usual in that context is the so-called genetic fallacy. To begin with, let us characterize the genetic fallacy. Since this paper is concerned with philosophical argumentation, it will be understood as an argumentative move aimed at supporting a point of view. However, the genetic fallacy could be described in several ways. One may, for instance, conceive it as a defective genetic explanation of the origin of a phenomenon (see Goudge, 1961). Without a doubt, this kind of explanation is important in science and other fields, and the problems founded in establishing the origin of a phenomenon, or a belief, could certainly affect the genetic argumentation in the sense we are considering. However, the argumentation which is usually considered as a case of genetic fallacy is one in which a genetic claim—an assertion according to which the genesis of an idea or phenomenon is such and such-, is used with the intention of stating something about its acceptability and differs, in that way, from genetic explanations. However, as observed with other fallacies, characterizing the genetic fallacy is no easy task. According to Douglas Walton, a genetic fallacy could be described as: The tactic of attacking the origin or original context of some practice, concept or argument that is instrumental to an opponent’s point of view, in an attempt to discredit that point of view unfairly. For example: “The wedding ring originated from the ankle chain used by men to confine their wives; therefore, wearing a wedding ring is a sexist (bad) practice.” (2010, p. 393)

5 Argument Evaluation in Philosophy: Fallacies …

67

In this characterization, the fallacy is committed in a context in which a viewpoint is criticized. For that reason, and provided that people are at least an important factor in the original context of ideas, this fallacy has been sometimes classified as a type of ad Hominem whereby a point of view is rejected on the basis of some negative trait of the person putting it forward (Cf. Copi, 1980, p. 84; Walton, 1998, p. 18). However, this kind of argumentation could be used not only to question a point of view but also to support it, and other factors apart from the person who originated the idea, could be considered. Salmon, in his Logic (1973), characterizes the genetic fallacy in a more general way. He states that, “The error of treating points of the discovery context as if they belonged to the context of justification is called ‘genetic fallacy.’” (1973, p. 11). According to this definition, the genetic fallacy could be committed in criticizing as well as in justifying a point of view and it is an infringement of the Relevance rule: the origin of an idea, its discovery and its history, are irrelevant with respect to its truth or falsehood. It is a significant fact that Salmon uses the distinction between context of discovery and context of justification. As it is well-known, Reichenbach’s distinction was a fundamental distinction in the context of twentieth century logic and philosophy of science. However, this distinction has had its detractors. In the context of the so called “historical turn” of philosophy of science, the exhaustiveness of that distinction, as well as its uses, have been deeply questioned. It is noteworthy that Salmon’s characterization of the fallacy is problematic, inasmuch as it seems to be committed to a particular philosophical position, which would undermine its potential as a philosophical evaluative strategy. In a more neutral way, the genetic fallacy could be understood, as Walton does, as the use of the genesis or history of a point of view, as a reason to prove or disprove it, a seemingly wrong move due to lack of relevance. This definition faces another difficulty: as Walton mentions, “arguments based on origins are not always fallacious.” (ibid.) For instance, his own example can be modified to thereby lose its fallacious character: (1) The wedding ring originated from the ankle chain used by men to confine their wives; for that reason, wearing a wedding ring could be a sexist (bad) practice and should be revised. Here, a practice is undermined or dismissed due to its origin. Still, it remains unclear why this should be understood as fallacious. The origin of a practice or an idea, its history and its function in specific contexts can be, and perhaps should be, something to be taken into account when being assessed. Otherwise, it seems that a priori principles, or God’s will, would be the only resource to assess a practice from a moral point of view.2

2 In the history of philosophical ethics, another possibility was considered, namely, the consequences.

However, the consideration of consequences as a relevant item in assessing moral principles was also dismissed as a fallacy, as an ad consequentiam fallacy (see Walton, 1999).

68

F. E. López

This problem has been noticed and attempts have been made to formulate criteria for distinguishing the genetic fallacy from correct uses of genetic argumentation (Crouch, 1993; Ward, 2010). In a recent article, Boudry, Paglieri, and Pigliucci (2015) address this problem in a general way and consider the genetic fallacy among others. In their opinion, even current pragmatic and dialectical approaches of fallacies fail in giving a proper account of the problem of the demarcation between fallacious and non-fallacious arguments. Furthermore, they maintain that those approaches either characterize fallacies in a simplistic, non-realistic, and non-persuasive way, or in a more sophisticated way which makes it almost impossible to distinguish between fallacious and non-fallacious arguments. As a consequence of this destructive dilemma, they argue in favor of “the intractable nature of fallacies” (p. 433). Although it is not clear whether this conclusion can be extended to all fallacies, it does seem to apply to the case analyzed here. At this point, it is possible to wonder whether the very idea of genetic fallacy should be completely rejected. It is certainly a possibility. Nevertheless, the concept of fallacy, and the genetic fallacy in particular, play an important part in philosophical controversies and sometimes lead to clearer argumentations. In this vein, Boudry, Paglieri and Pigliucci suggest that, instead of being wielded like the sword of judgment against opponents, [the concept of fallacy could be] employed for more amicable purposes, e.g., suggesting ways to clarify arguments that, without being necessarily flawed, stand in need of substantial elaboration (2015, p. 452).

This is an interesting suggestion which points to the fact that, in general, fallacies are seen as logical or communicative neutral errors in argumentation and, consequently, when a fallacy is identified in someone’s argumentation, it looks like the discussion has reached its end. Indeed, this could be the reason why, in a highly controversial field as philosophy, the charge of committing a fallacy is so common: it seems to provide a neutral departure point that must be shared by all the parties engaged in the argumentation. In that way, the allegedly logical, and consequently neutral, character of the error contained in a fallacy is the reason why a fallacy can be wielded as the sword of judgment. This happens even if the reasons why it should be considered as a fallacy are not clear. There, in its seeming neutrality, resides its persuasiveness. The offenders have now the burden of proof: they must explain themselves and show that their argument is not an instance of that fallacy or they must retract. In other words, this charge has an important rhetorical function which is accomplish by virtue of the use of a normative dialectical concept, the concept of fallacy. For that reason, the idea of strategic maneuvering, which introduces a rhetorical dimension without neglecting the dialectical normative facet of argument evaluation, provides some fruitful conceptual tools for understanding the charge of committing a fallacy.

5 Argument Evaluation in Philosophy: Fallacies …

69

5.3.2 The Genetic Fallacy in Question In order to better understand the charge of committing a genetic fallacy and its aftermath, it is worth looking into what philosophers have written about it. As mentioned before, around the middle of the twentieth century, in the context of North American philosophy, controversy arouse on the topic and different authors analyzed it. Since then, most of the articles which address the issue are intended to defend a philosophical point of view against what is understood as an attempt to block some philosophical research, such as naturalist ethical stances (Kahane, 2011), feminist studies (Crouch, 1991, 1993), Freudian studies (Kleiman, 1970; Pashman, 1970) or genealogical Nietzschean critiques (Kim, 1990), to mention a few examples. These replies share a similar strategy: instead of showing that their arguments cannot be considered as an instance of the genetic fallacy—thus accepting the very idea of genetic fallacy—they maintain that such form of argumentation need not be a fallacy and they question the assumptions upon which the charge of committing this fallacy is made (see Crouch, 1991; Feuer, 1983; Klement, 2002). The expression “genetic fallacy” in the context of American philosophy traces back to the 1920s and 1930s, and even earlier. Its first occurrence in a logic textbook is in Cohen and Nagel’s 1934 book—An introduction to Logic and Scientific Method— but it was probably a common expression before that date. Indeed, John Dewey refers to the genetic fallacy in 1928, in a way that assumes that it was a well-known concept. The expression also appears in an article by Cohen, “History Versus Value”, published in 1914. In that context, pragmatism, and specially Dewey’s version of it, was an influential philosophy in the United States. This is an important fact, provided that Dewey had been insisting, since the beginning of the twentieth century (see Dewey, 1902a, 1902b), on the need to use what he called the “genetic or historical method” in ethics and both Nagel and Cohen were familiar with, and critical of, Dewey’s work. In fact, in Cohen’s aforementioned article, the expression genetic fallacy is used in a footnote in which a position similar to Dewey’s is vehemently objected. Consequently, it is no far-fetched to assume that pragmatist ethics was one of the first targets of the proponents or inventors of the genetic fallacy and that consequently this fallacy was, at least in the beginning, more a partisan strategy than a neutral resource. This partisan character of the genetic fallacy is also suggested by the fact that this fallacy is more commonly used by philosophers in argumentative exchanges than by logicians. Indeed, the concept of genetic fallacy does not frequently appear in logic text books, while the idea of there being something like a genetic fallacy is frequent in the context of philosophy lectures, especially in the United States and primarily in those contexts where analytical philosophy is the prevailing tradition. This is what can be inferred from the articles in which the fallacy is analyzed. For example, in an article by Thelma Lavine we read: “Detection and censure of the genetic fallacy is one of the most securely established of contemporary philosophic practices.” (1962, p. 321). In the same vein, in a 1959 article it is stated that “Frequent references to the genetic fallacy are encountered in philosophic literature” and that “naturalists,

70

F. E. López

pragmatists and materialists are thought to be the chief offenders.” (Handy, 1959, p. 25). More recently, Crouch (1993, p. 227) and Klement (2002, p. 383) refer to the importance of fallacies and the genetic fallacy in philosophy learning environments. In these articles and others where this fallacy is addressed, different forms of genetic argumentation are defended against a seemingly unjustified attack trying to invalidate and block a whole philosophical approach. For example, Lavine (1962) criticizes how the genetic fallacy is used to invalidate contextual approaches within social sciences. Moreover, referring explicitly to Dewey, she states that the genesis of beliefs and valuations has a bearing on their validity. On his part, Handy suggests that It may be suspected that some writers are opposed to a scientific and developmental approach to ethics, and that this has led them to raise charges of the genetic fallacy in order to undermine such endeavors. (1959, p. 32)

Finally, Norwood Russell Hanson, in a 1967 article, entitled “The Genetic Fallacy Revisited,” questions the use of the genetic fallacy aimed at keeping history and philosophy of science apart. Here, the idea of there being a genetic fallacy is treated as equivalent to the dichotomy between context of discovery and context of justification. So far, three main conclusions can be established as regards this controversy. In the first place, it seems that the concept of genetic fallacy was, and still remains, an available resource for philosophers. Second, the genetic fallacy was mainly a criticism against specific philosophical stances, pragmatist ethics being one of the most important in the context analyzed. Finally, the defense of some genetic argumentation advocates was to consider the genetic fallacy not as an objection to a specific argument or move but as an attack against the entire approach they adopt. In this sense, the fallacy was not understood as pointing at a logical or argumentative error, but as a partisan strategy which already assumed the invalidity of the opponent’s view. This is particularly clear regarding philosophical ethics, since those who use the fallacy as a criticism are usually committed to aprioristic and transcendentalist conceptions of moral values. What has been said does not imply that there are no actual cases of arguments or argumentative moves which can be considered instances of the genetic fallacy. However, the fact that, despite the several rejections of the very idea of genetic fallacy, it continues to be an available resource suggests that this fallacy fulfils some important dialogical function, a function that could be shared with other fallacies. In order to better understand that function, in the next section, the charge of committing a fallacy is analyzed as a case of strategic maneuvering.

5.4 The Charge of Committing a Fallacy as Strategic Maneuvering In this section, the charge of committing a genetic fallacy is described as strategic maneuvering following the analytical model presented by van Eemeren (2010). In incorporating dialectical as well as rhetorical considerations, the concept of strategic

5 Argument Evaluation in Philosophy: Fallacies …

71

maneuvering is useful to explain the frequent appearance of fallacies in general, and of the genetic fallacy in particular, in the context of philosophical discussions. In a general way, strategic maneuvering “refers to the continual efforts made in all moves that are carried out in argumentative discourse to keep the balance between reasonableness and effectiveness” (2010, p. 40). This means that arguers try to have their point of view accepted, that is to say, they try to be effective in convincing the other party,3 but in a way that maintains their own image as “people who play the resolution game by the rules.” (2010, p. 42) This characterization perfectly matches with the charge of committing a fallacy, to the extent that it makes use of an alleged normative concept to point that the opponent’s move disregards a reasonableness rule. If the charge is fair, then the move is also functional with respect to the rhetorical aims of its proponent. Using the pragma-dialectical theoretical frame (see van Eemeren, 2010), the maneuver can be described as follows. Firstly, it is a move typically performed by the antagonist, that is to say, by the participant who expresses critical doubts regarding the protagonist’s point of view. In the case of the charge of committing a genetic fallacy, this implies that the antagonist is working only with the protagonist’s commitments and the rules for critical discussions. A fallacy is a logical or argumentative mistake the protagonist has made, and the antagonist restricts herself/himself to pointing out the error, and explaining it, if necessary. In other words, as antagonist, the participant who makes the charge is not supporting an opposite point of view. Secondly, the charge of committing a genetic fallacy is a move performed in the argumentation stage. According to van Eemeren, the dialectical and rhetorical aims of this stage are, respectively, the following: “To achieve clarity concerning the protagonist’s argumentation in defense of the standpoints at issue and the antagonist’s doubts concerning these standpoints and the argumentation in their defense” and “To establish argumentation that constitute an optimal defense of the standpoints at issue (by the protagonist) or to establish critical doubts that constitutes an optimal attack on the standpoints and the argumentation (by the antagonist)” (2010, p. 45). In this way, the charge of committing a fallacy would be dialectically useful to clarify the antagonist’s doubts and rhetorically convenient to the extent that it is an optimal attack which reveals a logical or argumentative error. Considering the three aspects of strategic maneuvering presented by van Eemeren, it can be said that the charge of committing a fallacy involves the following choices: 1. As regards topical potential, it can be said that the different approaches and classifications of fallacies offer the antagonist a large range of counter-arguments that can be used to clear up doubts and to establish objections. It might even be said that the reason why fallacies have been so popular among philosophers and logicians rests on the great topical potential that they have. In a field such as 3 The idea of effectiveness is however broader than the idea of persuasiveness: “the term effectiveness

is not exclusively applicable to argumentative moves made in the argumentation stage (as at least the term persuasiveness is), but also to argumentative moves made in the other dialectical discussion stages (which are not aimed directly at gaining acceptance of a standpoint)” (van Eemeren, Garssen, & Meuffels, 2012, pp. 51–52)

72

F. E. López

philosophy, in which every philosopher tends to start all over again, there is not much ready-made resources which can be used by philosophers belonging to different traditions. In that context, appealing to a well-known fallacy is certainly a promising strategic maneuver. 2. Regarding moves adaptation to meet the audience’s demand, the idea of fallacy is particularly suitable. Indeed, the period characterized by Frankena as a moment of expansion of the practice of labelling the opponent’s views as fallacies coincides with a moment of fruitful development of logic. Logical positivism and thereafter analytical philosophy were characterized by holding logic in high esteem, considering it as philosophically neutral and, what is more, as an unavoidable tool for overcoming metaphysical issues and improving philosophy. The importance of fallacies, and of the genetic fallacy in particular, in philosophy teaching environments makes the charge of committing one a very effective move in accordance with the audience’s demand. On the other hand, the idea that genesis and validity should stand as completely different issues, thoroughly extended in those contexts, further accounts for the importance of the genetic fallacy in particular. The references to the practice of detecting a genetic fallacy in philosophical teaching environments—that have been observed in the articles mentioned in Sect. 5.3—suggest that in some way, this type of fallacies worked as exemplars, in Kuhn’s sense. This could reinforce the idea that analytical philosophy can be understood as “normal philosophy”, as philosophy within a paradigm (see Levi, 2003), and could also explain the existence of a relatively homogenous audience, a fact important from a rhetorical point of view. 3. Finally, the third aspect of strategic maneuvering is “the exploitation of ‘presentational devices,’ which involves a choice as to how the argumentative moves are to be presented in the way that is strategically best” (van Eemeren, 2010, p. 94). As said, the idea of logic as an impartial starting point, and of fallacies as logical errors, makes them very effective presentational devices. This neutral character is also reinforced by the fact that fallacies are usually understood as procedural mistakes and not as mistaken premises or departure points. As suggested by Frankena in the epigraph of this paper, the different names and labels that philosophers use to refer to a fallacy plays an important role: in suggesting that someone has committed a well-known fallacy it seems that no more explanation is needed on behalf of the antagonist. She or he could simply ask: Are not you committing a genetic (or a naturalistic) fallacy? And the protagonist should now explain why this is not the case. As a consequence of this description, the charge of committing a fallacy and its importance in the mentioned philosophical contexts can be understood as a profitable strategic maneuvering which offers ready-made procedural and logical tools, as opposite to ad hoc stratagems, in a context in which logic was supposedly the keystone of philosophy and by means of a label that seems to do the whole work. Thus, the charge of committing a fallacy seems to be a masterstroke in philosophical critical discussions. And this could explain why, as Frankena suggests, philosophers are so inclined to invent fallacies.

5 Argument Evaluation in Philosophy: Fallacies …

73

Be that as it may, strategic maneuvering can derail. According to van Eemeren: The conditions that need to be fulfilled in order to ensure effectiveness do not necessarily always agree with the conditions that have to be met to guarantee reasonableness. (…) The pursuit of effectiveness may in some cases get the better of the simultaneous pursuit of reasonableness. Then the combination of effectiveness and reasonableness is out of balance. Making use of a railway metaphor, it can be concluded that the strategic maneuvering has derailed. (2010, p. 41)

This derailment is produced after some rule for critical discussion is violated. When such a thing happens, the move must be considered fallacious despite its effectiveness. For genetic argumentation advocates, some problems arouse in relation with the charge of committing a genetic fallacy. As seen, they have argued that the charge was indeed a baseless charge. Let us see now in which ways the analyzed maneuver could derail. In the first place, the analysis of the advocates of genetic argumentation’s replies suggests that they consider that, what is a seemingly argumentative maneuvering, that is, a maneuvering within the argumentation stage, is indeed a derailed confrontational maneuvering violating the Freedom Rule (Rule 1, “Discussants may not prevent each other from advancing standpoints or from calling standpoints into question”). In other words, it was seen as an attempt to block an entire approach, which would be based on a logical mistake. In other words, in such case the genetic fallacy is wrongly applied since it is not used against a particular advanced argument in the argumentation stage, in which the protagonist is trying to deduce the acceptability or unacceptability of a point of view based on the origin of the idea. Instead, the fallacy is used against a stance presented in the confrontation stage, independently of the arguments that could be presented in their favor. On the other hand, the charge of committing a genetic fallacy is derailed or fallacious when it violates the Starting Point Rule (Rule 6, “Discussants may not falsely present something as an accepted starting point or falsely deny that something is an accepted starting point”). As it is usually understood by most of the authors mentioned in Sect. 5.3, the genetic fallacy depends on a previously assumed premise according to which genesis and validity are mutually irrelevant. However, the approaches they adopt try to show how the genesis and the history of a belief can throw light on its validity. Consequently, the genetic fallacy cannot be considered as a shared procedural tool. It is, from this point of view, a partisan move which presents a starting point as if it were already accepted by the other party. Finally, it is worth mentioning another way in which the charge considered can derail. As mentioned before, the examples in which the genetic argumentation is held to be fallacious are usually cases of deductive argumentation which turn out to be defective. However, the genetic argumentation used in different contexts is usually a reasonable non-deductive move or argument based on symptomatic relations (see van Eemeren & Henkemans, 2017, p. 84). The origin of the idea or practice is treated as a symptom not of its falsehood or unacceptability, but of the need of revising it. The following quote illustrates the kind of argumentation that the advocates of genetic argumentation uphold. In his “Theory of valuation”, John Dewey asserts:

74

F. E. López Suppose, for example, that it be ascertained that a particular set of current valuations have, as their antecedent historical conditions, the interest of a small group or special class in maintaining certain exclusive privileges and advantages, and that this maintenance has the effect of limiting both the range of the desires of others and their capacity to actualize them. Is it not obvious that this knowledge of conditions and consequences would surely lead to revaluation of the desires and ends that had been assumed to be authoritative sources of valuation? (1988, p. 244)

Here, the fact that an idea or practice, in Dewey’s words a valuation, originated in a context in which it fulfils certain functions, along with the consequences of maintain this idea, are treated as a reason to reevaluate it. In this way, the charge of committing a genetic fallacy can indeed be a case of the straw man fallacy (violation of Standpoint rule 3, “Attacks on standpoints may not bear on a standpoint that has not actually been put forward by the other party”). That is the case when moves such as Dewey’s are interpreted as trying to deduce, in the strict sense of the word, something about the truth or falsehood of an idea. This is an important point because deductivism as an evaluative and interpretative strategy (see, Godden, 2005; Groarke, 1999) is generally the ground on which the charge of committing a genetic fallacy and other similar fallacies is made, at least in the context of philosophical critical discussions. It is worth mentioning that similar interpretations concerning other philosophical fallacies could be offered (López, 2015, 2018). For instance, the naturalistic and the ad consequentiam fallacy have received similar treatments. Strikingly, those fallacies have also been used against naturalistic and pragmatist stances on ethics in the same context that the genetic fallacy has, usually in derailed ways. Indeed, they could altogether be seen as a derailed discussion strategy, that is to say, as coordinated modes of strategic maneuvering “that are methodically designed to influence the result of a particular stage of the resolution process, or the discussion as a whole” (van Eemeren, 2010, p. 47). More specifically, it was, and it continues being, the discussion strategy, par excellence, of aprioristic and transcendentalist, and in some cases even emotivist philosophers, against non-transcendental fallibilistic and cognitivist stances on ethics.

5.5 Conclusion In this paper the genetic fallacy has been characterized and its use in philosophical discussions has been interpreted as strategic maneuvering. Such grasp is indeed an attempt to understand the way in which genetic argumentation advocates have dealt with the charge of committing that fallacy. Besides, some ways in which that maneuver can derail have been stated. It is important to mention that the analysis carried out applies to the genetic fallacy and other similar fallacies which are common in philosophy. Of course, the charge of committing a fallacy, as every move performed in a critical discussion, should be always understood as strategic maneuvering. The move could be understood as a demand for clearing some confusing point or as claim that the protagonist’s move is defective. However, in specific cases such as the ones

5 Argument Evaluation in Philosophy: Fallacies …

75

mentioned in this paper, the characterization of the move as fallacy assumes certain philosophical ideas which could be controversial and because of that, their value as evaluative strategies in philosophy is limited. An interesting point to emphasize is the value of the concept of strategic maneuvering as a tool in the analysis of philosophical argumentation. It is a remarkable fact that philosophers and especially analytical philosophers tend to assume very strong criteria to decide if an argument is cogent or not: prove or perish seems to be their motto. In other words, deductivism is the prevailing theory of argument evaluation among them. This is clearly seen in the uses of fallacies outlined before. For instance, the naturalistic fallacy is often presented in a deductive jargon: an “ought” cannot be derived from an “is.” However, since its revival in the twentieth century, argumentation theory has tried to broaden those criteria and with them, the very concept of rationality. The idea that charging someone of committing a fallacy is a strategic move that must combine effectiveness with reasonableness, and not a final word, could serve to enrich the philosophical practice of identifying and analyzing fallacies since it throws light on other functions of them, as suggested by Boudry, Paglieri and Piglucci. To remember that, as observed in all forms of argumentation aimed at supporting a point of view, reasonableness but also effectiveness and not only truth are the aims of philosophical controversialist, could help us to have more honest and constructive dialogues and to be careful to not let effectiveness prevail. One last point to emphasize is that this problem with the use of fallacies in philosophical critical discussion can be seen as an instance of a more general trait of philosophy. As it is commonly assumed within the pragma-dialectical theory of argumentation, the ideal model of critical discussion is based on a philosophy of reasonableness. According to van Eemeren, such philosophy “takes the fallibility of human reason as its starting point and elevates systematic critical testing to the guiding principle of problem solving by means of argumentation in all areas of human communication and interaction” (van Eemeren, 2010, p. 32). This explains a striking feature of philosophical discussions: as the very idea of reasonableness is usually—if not always as Habermas claims (1992, p. 1)—its subject-matter, or at least a point at issue, those discussions tend to be meta-discussions in van Eemeren’s sense, that is to say, discussions about the procedural starting-points that should be accepted by the parties (2010, pp. 241–242).

References Boudry, M., Paglieri, F., & Pigliucci, M. (2015). The fake, the flimsy, and the fallacious: Demarcating arguments in real life. Argumentation, 29, 431–456. Cohen, M. R. (1914). History versus value. Journal of Philosophy, Psychology and Scientific Methods, 11, 701–716. Cohen, M. E., & Nagel, E. (1934). An introduction to logic and scientific method. New York: Harcourt, Brace and Company. Copi, I. (1980). Introducción a la Lógica. Buenos Aires: Eudeba.

76

F. E. López

Crouch, M. A. (1991). Feminist philosophy and the genetic fallacy. Hypatia, 6, 104–117. Crouch, M. A. (1993). A limited defense of the genetic fallacy. Metaphilosophy, 24, 227–240. Dewey, J. (1902a). The evolutionary method as applied to morality. Philosophical Review, 11, 107–124. Dewey, J. (1902b). The evolutionary method as applied to morality: II. Its significance for conduct. Philosophical Review, 11, 353–371. Dewey, J. (1928). The inclusive philosophic idea. Monist, 38, 161–177. Dewey, J. (1988). Theory of valuation. In J. Dewey (Ed.), The collected works of John Dewey (Vol. 13, pp. 190–253). Carbondale: Southern Illinois University Press. Feuer, L. S. (1983). The genetic fallacy re-examined. In P. Kurtz (Ed.), Sidney Hook: Philosopher of democracy and humanism (pp. 227–246). Buffalo: Prometheus Books. Frankena, W. K. (1939). The naturalistic fallacy. Mind, 48, 464–477. Godden, David M. (2005). Deductivism as an interpretive strategy: A reply to Groarke’s recent defense of reconstructive deductivism. Argumentation and advocacy, 41, 168–183. Goudge, T. A. (1961). The genetic fallacy. Synthese, 13, 41–48. Groarke, L. (1999). Deductivism within pragma-dialectics. Argumentation, 13, 1–16. Habermas, J. (1992). Teoría de la acción comunicativa. Madrid: Taurus. Handy, R. (1959). The genetic fallacy and naturalistic ethics. Inquiry: An Interdisciplinary Journal of Philosophy, 2, 25–33. Hanson, N. R. (1967). The genetic fallacy revisited. American Philosophical Quarterly, 4, 101–113. Kahane, G. (2011). Evolutionary debunking arguments. Noûs, 45, 103–125. Kim, C. T. (1990). A critique of genealogies. Metaphilosophy, 21, 391–404. Kleiman, L. (1970). Pashman on freud and the genetic fallacy. Southern Journal of Philosophy, 8, 63–65. Klement, K. C. (2002). When is genetic reasoning not fallacious? Argumentation, 16, 383–400. Lakatos, I. (1974). The role of crucial experiments in science. Studies in History and Philosophy of Science Part A, 4, 309–325. Lavine, T. Z. (1962). Reflections on the genetic fallacy. Social Research, 29, 321–336. Levy, N. (2003). Analytic and continental philosophy: Explaining the differences. Metaphilosophy, 34, 284–304. López, F. E. (2015). Filosofía y falacias: el caso de la argumentación por las consecuencias en la filosofía pragmatista. X Jornadas de Investigación del Departamento de Filosofía, FaHCE UNLP. López, F. E. (2018). Ética y falacias: para no bloquear el camino de la investigación. Seminario Internacional de Estudios sobre Discurso y Argumentación (SEDiAr), 14 al 16 de Marzo. Instituto de Lingüística (FFyL, UBA) y Universidad Estaual de Santa Cruz (UESC). Inédito. Monk, R. (1996). Bertrand Russell: The spirit of solitude. New York: The Free Press. Pashman, J. (1970). Is the genetic fallacy a fallacy? Southern Journal of Philosophy, 8, 57–62. Perelman, C. (1979). The new rhetoric and the humanities: Essays on rhetoric and its applications. Dordrecht: Reidel. Perelman, C., & Olbrechts-Tyteca, L. (1969). The new rhetoric. A treatise on argumentation (Trans.). Notre Dame: University of Notre Dame Press. (Original work published in 1958). Roll-Hansen, N. (1989). The crucial experiment of Wilhelm Johannsen. Biology and Philosophy, 4, 303–329. Salmon, W. C. (1973). Logic. Englewood Cliffs: Prentice-Hall. van Eemeren, F. H. (2010). Strategic maneuvering in argumentative discourse, extending the pragma-dialectical theory of argumentation. Amsterdam and Philadelphia: John Benjamins Publishing Company. van Eemeren, F. H., Garssen, B., & Meuffels, B. (2012). Effectiveness through reasonableness: a pragmadialectical perspective: preliminary steps to pragma-dialectical effectiveness research. Argumentation, 26, 33–53. van Eemeren, F. H., & Snoeck Henkemans, F. (2017). Argumentation. Analysis and evaluation. New York and London: Routledge.

5 Argument Evaluation in Philosophy: Fallacies …

77

Walton, D. (1998). Ad Hominem arguments. Alabama: University Alabama Press. Walton, D. (1999). Historical origins of argumentum ad consequentiam. Argumentation, 13, 251–264. Walton, D. (2010). Genetic fallacy. In J. Dancy, E. Sosa, & M. Steup (Eds.), A companion to epistemology (2nd ed., p. 393). Singapore: Wiley-Blackwell. Ward, A. C. (2010). The value of genetic fallacies. Informal Logic, 30, 1–33.

Chapter 6

Dialogical Sequences, Argumentative Moves and Interrogative Burden of Proof in Philosophical Argumentation Joaquin Galindo

6.1 The Role of Questions in Argument In the process of argumentation, we not only provide reasons, but also ask questions. The questions we ask are used for several purposes. First, we use questions to express our doubts or disagreements (‘Is that really so?’, ‘Do you really think so?’) and to request reasons (‘Why do you claim that?’, ‘On what basis do you say that?’, ‘What is your evidence?’). Then, we use them to ask for clarification, either of the claim or of the argument itself or of its warrant (‘What do you mean?’, ‘What do the premises have to do with your claim?’, ‘How does that follow?’). All these are obvious uses, and it is undeniable that questions in one form or another play an important role in any argument (see van Eemeren, Houtlosser, & Snoeck Henkemans, 2007). However, this paper shall explore some properly argumentative functions of asking questions which, being less obvious, have received less theoretical attention. Briefly stated, my claim is that some argumentative moves are based on mechanisms of question generation that have the objective of allocating the burden of proof. Although these argumentative moves might occur in other domains, I shall argue that such argumentative moves are characteristically philosophical. But first, let me clarify what I understand by the key terms of “question-generating mechanisms” and “allocating the burden of proof”. It is a commonplace to distinguish between rhetorical and non-rhetorical questions. Rhetorical questions are ordinarily believed to be disguised claims, and so not really questions at all. It is probably more complicated than that, but I shall leave that issue aside in this paper. As for non-rhetorical—or authentic questions—they are very often considered to be requests for information. There is much evidence (Graesser, & Person, 1994) that this is an incomplete description, but in any case, we ought to distinguish at least four main question-generating mechanisms: J. Galindo (B) Department of Philosophy, University of Valladolid, Madrid, Spain e-mail: [email protected] © Springer Nature Switzerland AG 2020 F. H. van Eemeren and B. Garssen (eds.), From Argument Schemes to Argumentative Relations in the Wild, Argumentation Library 35, https://doi.org/10.1007/978-3-030-28367-4_6

79

80

J. Galindo

1. Knowledge deficit questions: These appear when the one who asks the question detects a lack of information in their knowledge base and questions are formulated in order to obtain the missing information. This is the standard use of questions as requests for information. (‘Who was present?’) 2. Common ground questions: They are formulated to make sure that a piece of knowledge is shared between the participants of the dialogue. (‘Would you agree that…?’) 3. Social coordination questions: They focus on the participants’ roles in the dialogue, such as in frequent student-teacher exchanges (‘Can we take the recess now?’). 4. Conversation control questions: These serve to direct the flow of the dialogue and/or the attention of the participants (‘May I ask you a question?’) All these mechanisms play a role in argumentative exchanges, but I am interested in common ground questions, particularly those that redirect the dialogue in counterargumentation in connection with knowledge deficit questions. The latter are usually raised when the questioner: 1. 2. 3. 4. 5.

Finds an obstacle or a problem in a plan. Encounters a contradiction. Perceives an unusual or abnormal event. Perceives an obvious lack in the questioner’s knowledge base. Finds a need to evaluate and make a decision between a set of alternatives that are equally likely.

I will focus on the role played by common ground questions and knowledge deficit questions in dialogical sequences in philosophical argumentation, i.e., argumentation that occurs in classical and contemporary works in the field of philosophy. The Common Ground Questions, by their function in philosophical argumentation, can be classified as: Questions that ask for clarification, which I will call ‘CGQ’, and dialectical compromise questions, which I will call ‘CGQ*’. On one hand, CGQs verify the understanding of a proposition by the other party, the interpretation of the thesis, serve to request examples or ask to elaborate a point. On the other hand, CGQ*s serve to make explicit dialectical commitments (Walton, 1998, p. 98), such as: – Propositions (statements that can be used as premises for arguments by the other party). – Arguments (generally, the logical minimum of the argument appears in the question). – Retractions or withdrawal of a thesis or other commitments (premises, warrants, counter-considerations). – Reformulations of the thesis or the direct answer. As for the Knowledge Deficit Questions that interest us in this paper, there are two types: the initial question, which I will call ‘KDQ’; and the auxiliary philosophical questions, which I will call ‘KDQϕ’ (Wi´sniewski, 1995, p. 171). Generally, a KDQ

6 Dialogical Sequences, Argumentative Moves and Interrogative …

81

initiates a dialogical sequence in philosophical argumentation. The auxiliary questions in philosophical argumentation usually attempt to make explicit principles of some kind. However, in this research we will illustrate two typical cases; the KDQϕs that serve to make explicit: the presuppositions of questions and argument’s warrants. There are many studies of questions in formal semantics and logic that are relevant to discussions about knowledge deficit and common ground questions (Which- and Why- questions, embedded and indirect questions, etc.). At the moment, the form of the questions is not of interest, but the central concept in all these, namely the presupposition of a question, is. Belnap and Steel (1976, p. 5) define a question as presupposing a statement if and only if the truth of the statement is a logically necessary condition for there being a true answer to the question. If an answer cancels the presupposition, then it is usually called a “corrective answer”. For example, the presupposition of the question: ‘When did the Third World War end?’, is: ‘There was a Third World War.’ If one answers that question by saying ‘It never happened’, or ‘There has not yet been a Third World War’, then one has given a corrective answer. As is well known, the many question fallacy is related to presuppositions (‘Have you stopped beating your wife?’, ‘Did you hide the stolen goods in your house?’). The only answer to such questions is a corrective answer which denies the presupposition. Recall its application in Walton’s Profile dialogues.1 In sum, the three auxiliary concepts for the part of my claim that regards mechanisms of question generation are: common ground questions (clarifying questions and dialectical questions), knowledge deficit questions (initial and auxiliary philosophical questions), and the presuppositions of a question (corrective answers).

6.2 Burden of Proof Now let me explain what I mean by “the allocation of burden of proof”. First of all, by “burden of proof” I mean the dialectic obligation often expressed in the norm: ‘The person one who makes a claim has the burden of proof.’ I follow Rescher’s distinction (Rescher, 1977, p. 27) between Initiating burden of proof (I-burden) and Evidential burden of proof (E-burden). The obligation to support an assertion with arguments in a dialectic situation has I-burden. On the other hand, counter-considerations and responses to counter-argumentation have E-burden. We can distinguish, with Marraud (2017, pp. 54–55) three types of counterargumentation each having a different E-burden. Using Toulmin’s well-known terminology to defining them, we have: • Objection: questions the truth of data, giving reasons. • Rebuttal: questions the warrant (Pollock’s ‘undercutting defeater’). 1 “The profile of dialogue represents a local sequence of moves that is one part of a longer sequence

of moves in a goal directed conversational exchange of a certain kind between two parties” (Walton, 1999b, p. 2).

82

J. Galindo

• Refutation: questions the claim by providing an argument with an opposite (contrary or contradictory) conclusion. The distribution of the burden of proof, both I-burden and E-burden, can be represented by the following tables (inspired by Lorenzen’s dialogical logic): Prop = Proponent Opp = Opponent W = Warrant C = Claim Rn = Reason * = Burden of proof ! = Assertion / = Because ¬ = negation Objection The proponent has the I-burden (Asher & Lascarides, 1998) and advances an argument (Baskent, 2016). If a counterargument challenges the reason given to hold the thesis, the E-burden is transferred to the opponent (Belnap & Steel, 1976) who discharges it by providing an argument (Bermejo-Luque, 2011) with the result of returning to the initial state (van Eemeren & Grootendorst, 1984) (Table 6.1). Rebuttal The proponent again has the I-burden (Asher & Lascarides, 1998) and discharges it with an argument (Baskent, 2016). If the warrant is challenged by the opponent, the E-Burden again falls on him (Belnap & Steel, 1976) and must in turn give a reason (Bermejo-Luque, 2011). The result is that the proponent’s argument is deleted and we return again to the initial state (van Eemeren & Grootendorst, 1984) (Table 6.2). Refutation The proponent begins with the I-burden (Asher & Lascarides, 1998) and discharges it (Baskent, 2016). In reply, the opponent denies the conclusion and the E-burden passes to the opponent (Belnap & Steel, 1976), who discharges it with an argument (Bermejo-Luque, 2011), which is, precisely, the refutation of the proponent’s claim. The result is that we have now two counter-oriented arguments. Deciding between Table 6.1 Objection

Prop

Opp

1. *!C 2. C/R1 3.

*¬R1

4.

¬R1 /R2

5. *!C

6 Dialogical Sequences, Argumentative Moves and Interrogative … Table 6.2 Rebuttal

Prop

83

Opp

1. *!C 2. C/w R1 3.

*¬/w

4.

*¬/w /R2

5. *!C

Table 6.3 Refutation

Prop

Opp

1. *!C 2. C/R1 3.

*!¬C

4.

¬C/R2

them requires weighing reasons. These counter-argumentation moves can be reiterated (Table 6.3). For my purposes, these examples and terminology provided above are enough to explain the meaning of the phrase “allocation of the burden of proof”.

6.3 Typical Moves in Philosophical Argumentation And now we come to the point—the argumentative moves, especially those having characteristics of philosophical argumentation. We find them in arguments, counterarguments, and even in argumentative plans and strategies. Some examples of patterns of argumentative moves in philosophy are: infinite regress, arguments to meaninglessness, spelling out of analogies, thought experiments, category mistake charge, transcendental arguments, aporetic groups, emptiness charge, and many more. What matters now is that the patterns of argumentative moves I have selected are paradigmatically philosophical, as found in Plato’s Dialogues to the most recent papers. I was tempted to call them “philosophical tricks”, not so much because they are, like tricks in general, everywhere, but because they take lay readers by surprise and strike them as strange, even awaking in them the suspicion of a sleight of hand. However, the word “trick” has negative connotations: it suggests a hoax. I don’t mean that, so I use the more neutral term “argumentative moves”. The position I hold is that the argumentative moves in philosophy have something in common with the destruction-tests by which engineers investigate the characteristics of materials (Ryle, 2009, p. 207). Certainly engineers stretch, twist, compress, and batter bits of metal until they collapse, but it is by just such tests that they determine what the metal can and can’t withstand. In somewhat the same way, philosophical argumentative moves assess the dialectical force of a complex network of

84

J. Galindo

concepts, distinctions, theses and arguments by fixing the precise forms of dialectical mishandling under which they refuse to work. Philosophy needs these dialectical tests because the traditional methods of philosophy are armchair ones. I propose the hypothesis that argumentative moves can be analyzed as dialogical sequences, such as in profile dialogue. The goals of these sequences are usually: – Find implicit commitments: assumptions, presumptions, ontological commitments and so on, – Changes in the allocation of the burden of proof, – Semantic ascent (meta-dialogue), – Introduction and evaluation of conceptual distinctions, – Reformulation of a thesis. – Assessment of the relevance of a distinction – Testing the consistency of the thesis. In this paper I will analyze three argumentative moves: – Revealing irrelevant distinctions – Analytic dilemmas – Pragmatic self-refutation. Below, I will explain each one and illustrate them with examples. But, before that, let us recall my original claim. The coarse-grained version was: Some argumentative movements are based on question-generating mechanisms that have the objective of allocating the burden of proof. A fine-grained version might be: Revealing irrelevant distinctions, analytic dilemmas and pragmatic self-refutation charges are based on common ground questions (clarifying questions and dialectical questions), knowledge deficit questions (initial and auxiliary questions), and the presupposition of a question (corrective answers), all of which aim to allocate I-burden and E-burden and others dialectical effects. I shall take them in turn:

6.4 Revealing Irrelevant Distinctions Philosophers are famous for making distinctions, finding ambiguities, carrying out semantic ascents, and analyzing the meanings of words. But what is much less well realized is that, in the case of philosophers, these activities are not merely lexicographical, but have an argumentative function. Now let me focus on a classical occupation of philosophers. To test a conceptual distinction with the help of which a question is answered is an old philosophical procedure that we find in Plato’s Parmenides (133a–165c), and in many contemporary philosophers. It is generally used against dualistic distinctions and has several ontological and epistemological versions, but the structure is always the same. Its dialectical sequence usually has two steps, with the mechanism being the following: first, the opponent asks a question, and then, through an exchange of questions

6 Dialogical Sequences, Argumentative Moves and Interrogative …

85

and answers (KDQϕ), the presuppositions of that question are made it explicit; second, the proponent advances a distinction to answer the initial question. The direct answer is the proponent’s thesis. The opponent seeks to show that the distinction is inconsistent with the presuppositions of the question, usually through a sequence of dialectical questions. The proponent grants platitudes and obvious consequences from the thesis (the proponent commits to these arguments). Finally, the opponent makes explicit the warrants of these arguments and challenges them (rebuttal). I will illustrate the second step of that mechanism with the following example, taken from Perry’s fictional dialogue: A Dialogue on Personal Identity and Immortality (1978). It’s a conversation between two characters: Gretchen Weirob, a teacher of Philosophy at a small Midwestern college, and her friend, Sam Miller, a chaplain. I will present the dialogical sequence in dialectical tables. In them, I will use the following symbolic notation (Table 6.4): * = Burden of proof Knowledge deficit questions ?KDQ = Initial question ?KDQϕ = Auxiliary philosophical question Common ground questions ? CGQ = clarifying questions ? CGQ* = dialectical questions Cn (n ≥ 1) = Commitments

The opponent asks the initial question (Is it logically possible for a person to survive his death?) And the presupposition of that question is numerical identity over time, and no qualitative identity allowed (ersatz) (Asher & Lascarides, 1998). The direct answer of the proponent was: ‘what survives is your mind or soul’ (Baskent, 2016). Through questions (?CGQ**), the mind-body distinction is introduced in epistemological terms (the soul cannot be seen or felt or touched or smelt, is immaterial) and the proponent makes explicit his commitment to that distinction (Belnap & Steel, 1976). Then, through dialectical questions she makes the proponent admit the compromise with an argument C3 (‘If we say that bodies at different times are the same person, it is because the two bodies had the same soul’) (van Eemeren & Grootendorst, 1984). She raises questions to make explicit the warrant of the argument C3 (Graesser et al., 1992; Hamblin, 1958; Harrell, 2016; Hintikka, 2007; Krabbe, 2006). She then raises dialectical questions to make sure and to make explicit the proponent’s commitment to the warrant and its interpretation a priori (Marraud, 2013, 2017; Passmore, 1961). Then she proceeds to a rebuttal of the warrant W, using the presupposition of the initial question (Perry, 1978). This argumentative move is represented in the Diagram 6.1. Now the proponent reformulates the warrant of the argument C3, guided by clarification questions and dialectical questions (Rescher, 1977; Ryle, 2009; Walton, 1989, 1991, 1992). The opponent presents another rebuttal using the distinction C1 and presupposition of the initial question (Walton, 1997). As shown in the Diagram 6.2.

86

J. Galindo

Table 6.4 Dialogical sequence revealing irrelevant distinctions Miller (Prop)

Gretchen (Opp)

1.

?KDQ You don’t have to convince me that survival is probable, for we both agree you would not get to first base. You have only to convince me that it is possible. The only condition is that it be real survival we are talking about, not some up-to-date ersatz survival, which simply amounts to what any ordinary person would call totally ceasing to exist

2. Your mind or soul is immaterial, lodged in your body while you are on earth. The two are intimately related but not identical. Now clearly, what concerns us in survival is your mind or soul (T*). It is this which must be identical to the person before me now, and to the one I expect to see in a thousand years in heaven 3.

? CGQ* So I am not really this body, but a soul or mind or spirit? And this soul cannot be seen or felt or touched or smelt? That is implied, I take it, by the fact that it is immaterial?

4. C1 (distinction) That’s right. Your soul sees and smells, but cannot be seen or smelt 5.

? CGQ* Now when you say I am the same person, if I understand you, that is not a remark about this body you see and could touch and I fear can smell. Rather it is a remark about a soul, which you cannot see or touch or smell. The fact that the same body that now lies in front of you on the bed was across the table from you at Dorsey’s—that would not mean that the same person was present on both occasions, if the same soul were not. And if, through some strange turn of events, the same soul were present on both occasions, but lodged in different bodies, then it would be the same person. Is that right?

6. C2 You have understood me perfectly. But surely, you understood all of this before! (continued)

6 Dialogical Sequences, Argumentative Moves and Interrogative …

87

Table 6.4 (continued) Miller (Prop)

Gretchen (Opp)

7.

? KDQϕ But wait. I can repeat it, but I’m not sure I understand it. If you cannot see or touch or in any way perceive my soul, what makes you think the one you are confronted with now is the very same soul you were confronted with at Dorsey’s? But how do you know you are talking to Gretchen Weirob at all, and not someone else, say Barbara Walters or even Mark Spitz!

8. Well, it’s just obvious. I can see who I am talking to 9.

? KDQϕ But all you can see is my body. You can see, perhaps, that the same body is before you now that was before you last week at Dorsey’s. But you have just said that Gretchen Weirob is not a body but a soul. In judging that the same person is before you now as was before you then, you must be making a judgment about souls—which, you said, cannot be seen or touched or smelt or tasted. And so, I repeat, how do you know?

10. Well, I can see that it is the same body before me now that was across the table at Dorsey’s. And I know that the same soul is connected with the body now that was connected with it before. That’s how I know it’s you. I see no difficulty in the matter 11.

? KDQϕ You reason on the principle, “Same body, same self.” (W)

12. C3 (W) Yes 13.

? KDQϕ And would you reason conversely also? If there were in this bed Barbara Walters’ body—that is, the body you see every night on the news—would you infer that it was not me, Gretchen Weirob, in the bed?

14. C4 Yes (continued)

88

J. Galindo

Table 6.4 (continued) Miller (Prop)

Gretchen (Opp)

15.

Rebuttal W KDQϕ But then merely extend this principle to Heaven, and you will see that your conception of survival is without sense. Surely this very body, which will be buried and as I must so often repeat, rot away, will not be in your Hereafter. Different body, different person. Or do you claim that a body can rot away on earth, and then still wind up somewhere else?

16. Reformulation No, I do not claim that. But I also do not extend a principle, found reliable on earth, to such a different situation as is represented by the Hereafter. That a correlation between bodies and souls has been found on earth does not make it inconceivable or impossible that they should separate. Principles found to work in one circumstance may not be assumed to work in vastly altered circumstances. January and snow go together here, and one would be a fool to expect otherwise. But the principle does not apply in southern California 17.

? CGQ* So the principle, “same body, same soul,” is a well-confirmed regularity, not something you know “a priori”

18. ? CGQ By “a priori” you philosophers mean something which can be known without observing what actually goes on in the world—as I can know that two plus two equals four just by thinking about numbers, and that no bachelors are married, just by thinking about the meaning of “bachelor”? 19.

Yes

20. C5 (W’) Then you are right. If it was part of the meaning of “same body” that wherever we have the same body we have the same soul, it would have to obtain universally, in Heaven as well as on earth. But I just claim it is a generalization we know by observation on earth, and it need not automatically extend to Heaven (continued)

6 Dialogical Sequences, Argumentative Moves and Interrogative …

89

Table 6.4 (continued) Miller (Prop)

Gretchen (Opp)

21.

Rebuttal W’ But where do you get this principle? It simply amounts to a correlation between being confronted with the same body and being confronted with the same soul. To establish such a correlation in the first place, surely one must have some other means of judging sameness of soul. You do not have such a means; your principle is without foundation; either you really do not know the person before you now is Gretchen Weirob, the very same person you lunched with at Dorsey’s, or what you do know has nothing to do with sameness of some immaterial soul. (Perry, 1978, pp. 2–5)

22. Return to 1 ?KDQ The presupposition of the survival question is numerical identity over time and the body is destroyed by death. Two bodies in deferential moments lodge the same soul.

so W makes no sense in the case of survival

Against

W: ‘Same body, same self’ a priori:

so Two bodies at different times are the same person.

Diagram 6.1 Rebuttal a priori distinction The presupposition of the survival question is numerical identity over time and Epistemological criterion of immateriality of the soul Two bodies in deferential moments lodge the same soul.

Therefore W’ is unknowable

against

W’ ‘Same body, same self’ a posteriori:

Therefore Two bodies at different times are the same person.

Diagram 6.2 Rebuttal a posteriori distinction

90

J. Galindo

The dialectical result is to return to the initial question, since the answer depended on an irrelevant distinction to answer the main question (Walton, 1998).

6.5 Analytic Dilemmas The next argumentative move is also characteristic of philosophy and consists of questioning the meaning of a word or the interpretation of a proposition with the objective of confronting the proposer with the following claim: If you mean such-and-such, that is the truth but trivial; if such-and-such, it is then not trivial, but neither is it obviously true. (Passmore, 1961, p. 14)

This is not an argumentation scheme, since there is no pattern of inference. Rather, there is a main question (about the meaning of a word or the interpretation of a statement), a series of auxiliary questions whose function is to make the opponent, in a dialectic situation, answer affirmatively by directing him to the analytic dilemma. There may be different versions depending on the second horn of the dilemma, for the first horn is always the same: 1. First horn: ‘What you say is trivial.’ Second horn: ‘What you say is not obvious’ (we do not know if it is true or false). 2. First horn: ‘What you say is trivial.’ Second horn: ‘What you say is prima facie false.’ 3. First horn: ‘What you say is trivial.’ Second horn: ‘What you say is patently false (or absurd).’ Each of these three combinations has consequences for the distribution of the burden of proof. Now let us consider an example in detail, excerpted from Williamson’s fictional dialogue: Tetralogue (2015). In spite of the fragmentary character of the text, it is a procedure, as I mentioned before, that we find repeatedly in the philosophical tradition. Two people are in a train discussing science and witchcraft. One is a confessed relativist; the other uses somewhat impertinent logic: Sarah: Let me try again. Can we express your relativism as ‘Every point of view is just a point of view’? Zac: In a sense, Sarah. We can talk that way if you like. Sarah: If Bob says ‘Witchcraft works’ and I reply ‘That’s just your point of view’, I’m refusing to endorse his statement. I’m accepting that it’s his belief but by putting in the word ‘just’ I’m rejecting the idea that it’s more than his belief, that witchcraft actually does work. In the same way, when Zac says something and then later says ‘That’s just my point of view’, he sounds as if he’s refusing to endorse his own earlier statement, disowning it. That’s the retreat Bob and I complained about. But now Zac tells us he isn’t disowning his earlier statements. Zac: Exactly, Sarah.

6 Dialogical Sequences, Argumentative Moves and Interrogative …

91

Sarah: You’re accepting that it’s what you believe but you’re not rejecting the idea that it’s more than that. You are adding to what you previously put on the table, not replacing it. So we should delete the word ‘just’, because you don’t really mean that it’s just your point of view. Zac: Delete it if you like, Sarah. The wording was yours, not mine. Sarah: I will. So instead of saying ‘Every point of view is just a point of view’, relativism says ‘Every point of view is a point of view’. Zac: So it is. Sarah: Yes, but that makes relativism utterly trivial! Everyone, even the most hard-line absolutist, will agree that every point of view is a point of view (Williamson, 2015, pp. 32–33).

Let’s look at the burden of proof table (Table 6.5): * = Burden of proof Knowledge deficit questions ?KDQ = Initial question ?KDQϕ = Auxiliary philosophical question Common ground questions ? CGQ = clarifying questions ? CGQ* = dialectical questions Cn (n ≥ 1) = Commitments Corrective Answer = CA. AD = Analytic dilemma.

Table 6.5 Dialogical sequence analytic dilemma Prop (Zac) 1.

Opp ? CGQ* Can we express your relativism as ‘Every point of view is just a point of view’?

2. Yes C1*

If Bob says ‘Witchcraft works’ and I reply ‘That’s just your point of view’, I’m refusing to endorse his statement. Warrant: If A says that p and B says that p is just your point of view, then (material inference) B does not endorse p.

so

“just your point of view” implies disowning the embedded proposition

Zac asserts “‘Every point of view is just a point of view’

so Zac says something and then later says ‘That’s just my point of view’, he sounds as if he’s refusing to endorse his own earlier statement, disowning it.

3. CA (not disowning the embedded proposition)

? CGQ* accepting that it’s what you believe but you’re not rejecting the idea that it’s more than that. You are adding to what you previously put on the table, not replacing it. So we should delete the word ‘just’, because you don’t really mean that it’s just your point of view.

4. Yes (you can delete “just”)

One horn analytic Dilemma: Yes, but that makes relativism utterly trivial! Everyone, even the most hard-line absolutist, will agree that every point of view is a point of view.

5.*I-burden

92

J. Galindo

6.6 Pragmatic Self-refutation Another characteristically philosophical procedure (from Plato to Nagel) is frequently used in discussions against relativists and sceptics, that is, finding inconsistencies, but of a specific kind. This procedure is used to show that someone held ‘p’ and, say at another time, held ‘not p’ (like a philosopher in two different papers or books), for we might call that a mere inconsistency, and the proponent can reply by saying that they have changed their mind. Nor is it a reduction in the sense of extracting remote consequences involving the self-contradictory statement ‘p and not p’. Rather, it looks like a pragmatic paradox, like saying “I can’t talk,” in the sense that something that is said is the best counterexample of what is stated. Of course, as we shall see, the procedure is more subtle. Again, we find a set of auxiliary questions that lead the opponent step by step until the desired end is reached by showing that what is done is the best counterexample to the content of their thesis. The only defensive movement or choice is to argue that it is not really a counterexample. Consider another example from Williamson’s book: Zac: Whoa there, everyone, not so fast! Surely once you start using the word ‘true’ you imply certainty. You can’t call something true unless it’s certain, beyond doubt. If you don’t want to imply certainty, don’t say ‘true’. Sarah: Let me ask you a question. You’ll see why in a minute. Is it true that there is life on other planets? Zac: I’ve no idea, Sarah. Nobody knows—at least, not from my point of view. Sarah: OK, but is it certain that there is life on other planets? Zac: No, Sarah, it isn’t. As I just said, nobody knows. Sarah: There you are. Truth doesn’t require certainty, even from your point of view. Zac: Sarah, what are you talking about? I get to tell you what my point of view is. You can tell me what your point of view is. You don’t get to tell me what my point of view is. Roxana: Listen and learn. I will analyse the logic of Sarah’s questions. If truth entails certainty, then uncertainty entails untruth. Yes? Zac: Yeah, OK, we can play those logic games if you insist, Roxana. Roxana: I insist. Sarah’s first question was ‘Is it true that there is life on other planets?’; call that the truth question. Her second question was ‘Is it certain that there is life on other planets?’; call that the certainty question. Your response to the truth question was ‘I don’t know’. Your response to the certainty question was ‘No’. Do you accept that those were your responses? Zac: Yeah, Roxana, they were, but where is all this heading? Roxana: You do not treat truth and certainty as the same. Your responses to the truth question and the certainty question were not equivalent. ‘I don’t know’ is quite different from ‘No’. Zac: They are both negative responses, Roxana. What’s the big difference supposed to be?

6 Dialogical Sequences, Argumentative Moves and Interrogative …

93

Table 6.6 Dialogical sequence pragmatic self-refutation Prop (Zac) 1. the word ‘true’ implies certainty ¨*

Opp ?KDQ Is it true that there is life on other planets?

2. Answer: I’ve no idea, Sarah. Nobody knows— at least, not from my point of view. 3. Answer: No, it isn’t. As I just said, nobody knows.

?KDQ OK, but is it certain that there is life on other planets?

PRAGMATIC SELF-REFUTATION: There you are. Truth doesn’t require certainty, even from your point of view * Suppose that X asserts that the word ‘true’ entails certainty.

If truth entails certainty, then uncertainty entails untruth. Therefore

If truth entails certainty, TQ and CQ have the same presupposition.

If X inferred untruth from uncertainty, X ought to answer ‘No’ to the truth question, since ‘No’ to the certainty question

‘Is it true that there is life on other planets?’ Zac answer: ‘I don’t know’ ‘Is it certain that there is life on other planets?’ Zac answer: No

Therefore In practice, Zac does not take truth to require certainty.

Sarah: When you next have to apply for a visa, try responding to the question ‘Have you ever been involved in smuggling drugs?’ with ‘I don’t know’ instead of ‘No’. You will soon find out the difference. Roxana: If you, Zac, inferred untruth from uncertainty, you would have answered ‘No’ to the truth question, since you answered ‘No’ to the certainty question. But you did not. Sarah’s questions caught you out. They revealed that in practice even you do not take truth to require certainty (Williamson, 2015, pp. 51–54).

Let’s look at the burden of proof table (Table 6.6): TQ = The truth question: Is it true that p? CQ = The certainty question: Is it certain that p? (Diagram 6.3).

6.7 Conclusions Table 6.7 details the initial situation and goals of each of the three previous dialogical sequences. The first thesis I am defending is that some argumentative movements, being characteristically philosophical, are based on question-generating mechanisms that

94 Suppose that X asserts that the word ‘true’ implies certainty.

J. Galindo If truth entails certainty, then uncertainty entails untruth. so

If truth entails certainty, TQ and CQ have the same presupposition.

‘Is it true that there is life on other planets?’ Zac’s If X inferred untruth from uncertainty, X ought to answer ‘No’ to the truth answer: ‘I don’t know’. ‘Is it certain that there is question, since ‘No’ to the certainty question life on other planets?’ Zac’s answer: No so In practice, Zac does not take truth to require certainty.

Diagram 6.3 Pragmatic self-refutation Table 6.7 The initial situation and goals of each of the three previous dialogical sequences Type of dialogical sequence

Initial situation

Participant’s goal

Goal of dialogical sequence

Revealing irrelevant distinctions

She asks a question. She makes explicit the presupposition of the question The answer to the question introduces a distinction.

Show whether a distinction is consistent with the presupposition of the question

Assessment of the relevance of a distinction Dialectical effect of the failure: they are taken back to the initial question

Analytic dilemmas

Proponent’s thesis without burden of proof (T)

Opp: sequence of questions about interpretations of the thesis to arrive at the dilemma: “If you mean such-and-such, that is the truth but trivial; if such-and-such, it is then not trivial, but neither is it obviously true”

To change burden of proof (E-burden to T)

Pragmatic self-refutation

Proponent’s thesis with I-burden

Opp: Show that holding the thesis is the best counterexample against it Prop: Show that this is not a true counterexample

To test consistency of the thesis Dialectical effect of the failure: retraction of thesis or reformulation

6 Dialogical Sequences, Argumentative Moves and Interrogative …

95

have the objective of allocating the burden of proof and various dialectical effects. My working hypothesis is that argumentative moves can be analyzed as dialogical sequences. A fine-grained version is: Revealing irrelevant distinctions, analytic dilemmas and pragmatic self-refutation charges are based on common ground questions (clarifying questions and dialectical questions), knowledge deficit questions (initial and auxiliary questions), and the presupposition of a question (corrective answers), all of which aim to allocate I-burden and E-burden and others dialectical effects. Three examples are analyzed in which techniques from Rescher‘s formal dialectic (1977), Walton’s profiles of dialogue (1999a) and Marraud’s diagrams of argumentation techniques (2013) are used with diagrams supporting my thesis. Although research is restricted to these examples of philosophical argumentation, the analysis techniques are easily extrapolated to other arguments. Acknowledgements This work has been made possible by a Ph.D. scholarship from CONACYT, scientific agency of the Mexican government.

References Asher, N., & Lascarides, A. (1998). Questions in dialogue. Linguistics and Philosophy, 21, 237–309. Baskent, C. (Ed.). (2016). Perspectives on interrogative models of inquiry. Developments in inquiry and questions. Dordrecht: Springer. Belnap, N. D., & Steel, T. P. (1976). The logic of Questions and Answers. New Haven: Yale University Press. Bermejo-Luque, L. (2011). Giving reasons: A linguistic–pragmatic approach to argumentation theory. Dordrecht: Springer. van Eemeren, F. H., & Grootendorst, R. (1984). Speech acts in argumentative discussions: A theoretical model for the analysis of discussions directed towards solving conflicts of opinion. Dordrecht: Foris. van Eemeren, F. H., Houtlosser, P., & Snoeck Henkemans, A. F. (2007). Argumentative indicators in discourse. A pragma-dialectical study. Dordrecht: Springer. Graesser, A. C., & Person, N. K. (1994). Question asking during tutoring. American Educational Research Journal, 31(1), 104–137. Graesser, A. C., Person, N. K., & Huber, J. D. (1992). Mechanisms that generate questions. In T. E. Lauer, E. Peacock, & A. C. Graesser (Eds.), Questions and information systems (pp. 167–187). Hillsdale: Lawrence Erlbaum Associates. Hamblin, Ch L. (1958). Questions. The Australasian Journal of Philosophy, 36, 159–168. Harrell, M. (2016). What is the argument? An introduction to philosophical argument analysis. Cambridge (MA) & London: The MIT Press. Hintikka, J. (2007). Socratic epistemology: Explorations of knowledge-seeking by questioning. Cambridge: Cambridge University Press. Krabbe, E. C. W. (2006). Dialogue logic. In D. M. Gabbay & J. Woods (Eds.), Handbook of the history of logic, 7. Logic and the modalities in the twentieth century (pp. 665–704). Amsterdam: Elsevier. Marraud, H. (2013). ¿Es lógic@? Análisis y evaluación de argumentos. Madrid: Editorial Cátedra. Marraud, H. (2017). Las siete maneras de argumentar. Quadripartita Ratio. Revista de Retórica y Argumentación, 4, 52–57.

96

J. Galindo

Passmore, J. (1961). Philosophical reasoning. New York: Charles Scribner’s Sons. Perry, J. (1978). A dialogue on personal identity and immortality. Indianapolis: Hackett Publishing. Rescher, N. (1977). Dialectics. A controversy-oriented approach to the theory of knowledge. Albany: State University of New York Press. Ryle, G. (2009). Collected essays 1929–1968. New York: Routledge. Walton, D. (1989). Question-reply argumentation. New York: Praeger. Walton, D. (1991). Critical faults and fallacies of questioning. Journal of Pragmatics, 15, 337–366. Walton, D. (1992). Questionable questions in question period: Prospects for an informal logic of parliamentary discourse. In E. M. Barth & E. C. W. Krabbe (Eds.), Logic and political culture (pp. 87–95). Amsterdam: North-Holland. Walton, D. (1997). Judging how heavily a question is loaded: A pragmatic method. Inquiry: Critical Thinking Across the Disciplines, 17(2), 53–71. Walton, D. (1998). The new dialectic: Conversational contexts of argument. Toronto: University Press. Walton, D. (1999a). The fallacy of many questions: On the notions of complexity, loadedness and unfair entrapment in interrogative theory. Argumentation, 13, 379–383. Walton, D. (1999b). Profiles of dialogue for evaluating arguments from ignorance. Argumentation, 13, 53–71. Williamson, T. (2015). Tetralogue. Oxford: Oxford University Press. Wi´sniewski, A. (1995). The posing of questions: Logical foundations of erotetic inferences. Dordrecht: Springer.

Chapter 7

Eudaimonistic Argumentation Andrew Aberdein

Virtue theories have lately enjoyed a modest vogue in the study of argumentation, echoing the success of more far-reaching programmes in ethics and epistemology. Virtue theories of argumentation (VTA) comprise several conceptually distinct projects, including the provision of normative foundations for argument evaluation and a renewed focus on the character of good arguers. One of the boldest of such projects is the pursuit of the fully satisfying argument, the argument that contributes to human flourishing. An analogous epistemic project has been pursued independently as eudaimonistic virtue epistemology. The two projects have much in common: both projects stress the importance of widening the range of cognitive goals beyond, respectively, cogency and knowledge; both projects emphasize social factors, the right sort of community being indispensable for the cultivation of intellectual virtues necessary to each project. This paper seeks to unite the two projects, by arguing that the intellectual good life sought by eudaimonistic virtue epistemologists is best realized through the articulation of an account of argumentation that contributes to human flourishing. In the first section I survey VTA, with particular attention to the treatment of the fully satisfying argument. In the second section, I turn to eudaimonistic virtue epistemology, and discuss how it may be extended to encompass VTA. In the third section, I address the specific issue of adversariality, arguing that the picture proposed in the previous section promises a resolution of a long-contested debate. Lastly, I conclude with a section briefly exploring the prospects for a combined eudaimonistic project.

A. Aberdein (B) School of Arts and Communication, Florida Institute of Technology, Melbourne, USA e-mail: [email protected] © Springer Nature Switzerland AG 2020 F. H. van Eemeren and B. Garssen (eds.), From Argument Schemes to Argumentative Relations in the Wild, Argumentation Library 35, https://doi.org/10.1007/978-3-030-28367-4_7

97

98

A. Aberdein

7.1 Virtue Theories of Argumentation Dan Cohen has said that the “core ideas” of VTA “can fit on a couple of bumper stickers” (Cohen, 2013, 482): For a Good Argument, Argue W ell

Arguing W ell Requir es Good Arguer s In other words, VTA involves two successive steps away from much orthodox argumentation theory: firstly, a shift from the investigation and analysis of arguments as texts, broadly construed, to an appraisal of arguments as processes with participants; secondly, a concentration on the character of the participants—an expectation “that the admirable conduct of arguers ought to stem from virtues, inculcated habits of mind, rather than be accidental or occasional manifestations” (ibid.). In many respects, the first of these steps is the more consequential: the emphasis on virtue is arguably a more modest move than the shift of focus onto arguers from arguments. For that matter, just as not all virtue ethicists follow Aristotle in linking virtues to eudaimonia, neither do proponents of VTA necessarily see any analogous connection. Nonetheless, this connection does suggest a possible third step for VTA: from the character of arguers to the pursuit of human flourishing through the optimal conduct of argument. Here are some of the virtues that Cohen focuses on: willingness to listen to others; willingness to modify one’s own position; willingness to question the obvious; willingness to engage in serious argumentation (Cohen, 2005, 64). He situates them as Aristotelian means: extremes of excess or deficiency on their various axes comprise the corresponding vices. In my own attempt at a list of argumentational virtues, I complicate Cohen’s picture with subheadings, including some much more familiar virtues, many of them drawn from what is known as responsibilist virtue epistemology1 : 1. willingness to engage in argumentation (a) being communicative (b) faith in reason (c) intellectual courage i. sense of duty 2. willingness to listen to others (a) intellectual empathy i. 1 Virtue

insight into persons

epistemologists are often distinguished between responsibilists, whose virtues mostly look like Aristotelian character virtues, and reliabilists, whose virtues are less obviously virtue-like: they are reliable cognitive processes (see, for example, Battaly, 2008, 644).

7 Eudaimonistic Argumentation

99

ii. insight into problems iii. insight into theories (b) Fairmindedness i.

justice

ii. fairness in evaluating the arguments of others iii. open-mindedness in collecting and appraising evidence (c) recognition of reliable authority (d) recognition of salient facts i. sensitivity to detail 3. willingness to modify one’s own position (a) common sense (b) intellectual candour (c) intellectual humility (d) intellectual integrity i.

honour

ii. responsibility iii. sincerity 4. willingness to question the obvious (a) appropriate respect for public opinion (b) autonomy (c) intellectual perseverance i.

diligence

ii. care iii. thoroughness (Aberdein, 2010, 175).

One way in which Cohen has framed his approach to VTA that differs from mine is that he asks what a maximally good argument would look like: “What would make an argument satisfying to the point that the participants could say at the end, ‘Now that was a good argument’?” (Cohen, 2013, 477). He stresses that “The notion of an argument that is fully satisfying to its participants includes more than epistemic gains. It has to at least extend to cognitive gains more broadly, including emotional, ethical, and possibly aesthetic aspects as well” (Cohen, 2013, 478). He also notes that, while the virtues of argument may be necessary for the fully satisfying argument, they are not sufficient: “There are necessary conditions beyond the scope of the behavior and characters of the arguers, including the subject matter, the context, and the personal chemistry of the arguers” (Cohen & Miller, 2016, 459). Other argumentation theorists, less directly associated with VTA, have also said some similar things. For example, Kathryn Norlock has proposed Nel Noddings’ “receptivity” as a virtue of argument somewhat analogous to Cohen’s willingness to listen, a mean between the uncritical acceptance of others’ views and listening only in a defensive and uncharitable manner (Norlock, 2014, 2). Saliently, Norlock observes that for Noddings receptivity “seems to be essential to living fully as a person” (Noddings, 1984, 35). As Norlock notes, this suggests “the necessity of

100

A. Aberdein

receptivity to eudaimonia” (ibid.). Moira Howes also asks “does happiness increase the objectivity of arguers”, happiness here being thought of as not just subjective, but also eudaimonic: Argumentative communities that support subjective and eudaimonic happiness will have an easier time strengthening these factors. In such contexts, arguers and audiences are likely to find it easier to change their minds in the face of evidence and imagine diverse, multifaceted audiences. They are more likely to be resilient and assertive in the face of negative commentary and find it easier to disagree respectfully. Overall, we can expect such communities to support a deeper commitment to acquiring an empirically adequate understanding of reality (Howes, 2014, 8).

So there are reasons to expect eudaimonia to have positive dividends in argumentational practice. Sharon Bailin, responding to Howes, introduces an important clarification: Eudaimonia, in the Aristotelian sense, comes in fulfilling our human purposes, i.e., living in accordance with reason. Thus, on this account, critical thinking (and hence objectivity) is constitutive of eudaimonic happiness. So it’s not so much that eudaimonia is a means to the end of objectivity in argumentation. It is rather that part of what it means to live well is to engage in rational pursuits and rational exchanges (Bailin, 2014, 3).

Arguing well is not just a ladder you climb up to get to a state of eudaimonia and may then throw away; it is actually an indispensable part of the goal you are seeking: critical thinking as a constitutive component of the good life. The version of VTA that Cohen and I have defended may be seen as a “virtues in” account: in other words, possession of various virtues might be necessary to being a good arguer. The converse of this account is implicit in Bailin’s idea that being a good arguer might be constitutive of what it is to lead the good life, and therefore that arguing badly could potentially corrupt and corrode your capacity to lead the good life. This idea has also been defended by Sherman Clark, a professor of law who has been writing about virtues in the context of legal persuasion for some years (Clark, 2003, 2011, 2013): the links between VTA and such closely related work in virtue jurisprudence are seldom addressed but hold great promise. Clark uses an example of an arguer with a perfectly legitimate goal of which they want to persuade some audience—his example concerns transport policy. Their view is perfectly respectable in itself, they’re not secret racists who are trying to convince people of racist things for racist reasons; they just have a point of view on transport policy. But they might cynically appreciate that, if they add a touch of racist dog whistle to their transport policy arguments, they could win over some other people who were less admirable than themselves.2 But that would still be a bad thing to do. It might be, in the short term at least, successful; they might actually win some people over. But even if it is a good argument in that very reductive sense, it is certainly not a virtuous argument. As Clark says, You might well choose to eschew that argument. Why? In part because you might consider it simply wrong to appeal to racism. But you might also quite sensibly realize that by making 2 For

a discussion of dog whistling in the context of VTA, see (Aberdein, 2018, 20).

7 Eudaimonistic Argumentation

101

that sort of argument, even and perhaps especially if you did so indirectly and subtly, you would not just be appealing to but also helping to construct and reinforce fear and prejudice (Clark, 2011, 852).

What makes that kind of argument bad is not just a pious sense that one shouldn’t be doing these kind of things, it’s that you’re actually making the world worse by indulging in such arguments and by encouraging other people to do likewise. That’s the point that Clark is making, drawing on the idea that argument plays a role in constructing character: The basic claim, however, is this: persuasion succeeds, if it succeeds, not by force of abstract logic, but by finding or making space in the constitutive world views—the beliefs, understanding, and priorities—of those we persuade. Unless we believe that the beliefs and understandings and priorities of those we persuade are always fully developed and permanently fixed, we should recognize that when we persuade we do not simply find, but, crucially, sometimes also may make space in their world views. When we persuade, we also teach; when we teach, we may construct. If so, we should think about what we may be doing to people when we persuade them (Clark, 2013, 1376).

This reverses the causal relationship between argument and character usually proposed in VTA. Clark is not just saying that you have to have the right sort of character in order to produce the right sort of arguments; he is saying you have to produce the right sort of arguments in order to have the right sort of character. When you succeed in persuading people of things, you succeed in some small way in shaping their character. If their character gets shaped by a lot of bad arguments then that will not lead anywhere good.

7.2 Intellectual Flourishing A very specific project within virtue epistemology is what Britt Brogaard has called “eudaimonistic epistemology” (Brogaard, 2014a, 2014b, 2014c). Her proposal is that we can do epistemology successfully with a fundamental norm that emphasizes intellectual flourishing, rather than the sorts of things that fundamental epistemic norms usually emphasize, such as truth: Truth Norm You ought to maximize your true beliefs and minimize your false beliefs (Brogaard, 2014c, 139).3

Brogaard’s point is not that we should throw away the truth norm; on the contrary, we shouldn’t say things that aren’t true, and we should expect our norms to require that. But the truth norm need not be the fundamental norm. For Brogaard, “the fundamental epistemic norm is not to hinder intellectual flourishing” (Brogaard, 2014a, 15). We should still endorse the truth norm, but because the truth norm is in general conducive to intellectual flourishing, not because it is a fundamental norm. 3 Brogaard

credits versions of this view to numerous epistemologists, including Ralph Wedgwood, Tyler Burge, and Christopher Peacocke.

102

A. Aberdein

This general principle can be cashed out in specific instances. Brogaard suggests the following: Intellectual Flourishing (belief) You should believe p only if believing p does not hinder intellectual flourishing. Intellectual Flourishing (assertion) You should assert p only if asserting p does not hinder intellectual flourishing. Intellectual Flourishing (action) You should treat p as a reason for action only if treating p as a reason for action does not hinder intellectual flourishing (Brogaard, 2014a, 15).

There is a much bigger story to be explored here as to whether or not this actually works. Brogaard has offered a full defence, as I shall not. She suggests some issues in epistemology, such as the value of knowledge, where treating the intellectual flourishing norm as more fundamental than the truth norm has clear advantages. Indeed, there are some very good reasons why we should not want to build our intellectual lives merely upon the acquisition of true beliefs. Were I to begin obsessively counting leaves I would acquire any number of true beliefs in short order, but that does not sound like any sort of good life (Brogaard, 2014a, 16). Since such behaviour would give me true beliefs at the cost of inhibiting any prospect of intellectual flourishing, it would be ruled in by the truth norm but ruled out by the intellectual flourishing norm. Just as Cohen observes that argumentational virtues are necessary but not sufficient for a fully satisfying argument, Brogaard also notes that virtues alone do not suffice for intellectual flourishing: “While virtuous character traits and well-functioning cognitive faculties and abilities can lead to a good intellectual life, there are many cases in which true belief flows from virtuous character traits or well-functioning cognitive faculties and abilities but in which the agent is not on the right track intellectually speaking” (Brogaard, 2014b, 97). In this respect, of course, both authors echo Aristotle (and differ from Plato or the Stoics) in treating eudaimonia as also dependent on other external goods, whose presence or absence may be a matter of luck (Hursthouse & Pettigrove, 2018, §2.1). Thus Brogaard observes that “Just as we cannot flourish, in Aristotle’s sense, in solitude, so we cannot flourish intellectually outside an intellectual community” (Brogaard, 2014c, 140). Likewise Cohen avers that “The wrong time or place can be as detrimental to the success of an argument as bad arguers. But so can the wrong arguers, no matter how good they are” (Cohen, 2013, 484). With only minor modifications, Brogaard’s norm of intellectual flourishing can produce norms for argumentation. Her specific instances of the norm for belief and assertion apply without modification to fundamental actions within argumentation: the acceptance of conclusions and advancing of premisses. Indeed, assuming practical reasoning is included within the scope of argumentation, then her action instance might also find an unmodified application. But Brogaard does not contend that the instances that she specifies exhaust the possible applications of her fundamental norm. It is entirely within the spirit of her programme to propose further instances. For example:

7 Eudaimonistic Argumentation

103

Intellectual Flourishing (reasons) You should treat p as a reason for accepting q only if treating p as a reason for accepting q does not hinder intellectual flourishing

Connoisseurs of recent debates on the normativity of logic may recognise this instance of the norm as at least an echo of what has come to be known as a “bridge principle” (MacFarlane, 2004). That is, it bridges the two independent realms of logic and of normativity. Specifically, it may be seen as a possible way of spelling out the consequent in a generic bridge principle of the form “If p ⇒ q, then (normative claim about believing p and q)” (cf. Dutilh Novaes, 2015, 591; MacFarlane, 2004, 6).4 However, since argumentation, at least according to VTA, is an essentially dialogical activity, it seems appropriate that the instances of the norm of intellectual flourishing should be expressed in terms of a relationship between agents. For example, where a (perhaps defeasible) inference of the form p ⇒ q obtains: Intellectual Flourishing (proponent) If Opponent has granted p, then Proponent should put forward q (and require Opponent to grant it) only if doing so does not hinder intellectual flourishing. Intellectual Flourishing (opponent) If Opponent has granted p, and Proponent puts forward q, then Opponent should either grant q or retract his endorsement of p, depending on which hinders intellectual flourishing the least.

These two instances are adapted from Catarina Dutilh Novaes’s work extending John MacFarlane’s account of bridge principles to dialogical contexts (Dutilh Novaes, 2015, 604 ff.). They may perhaps be seen as special cases of the reasons instance of the intellectual flourishing norm. Although this is the barest sketch of how one might apply a eudaimonistic epistemology to thinking about how arguments work, it should at least demonstrate that such an application appears worthy of further enquiry.

7.3 Adversariality As a test of this approach, I want to explore whether such an account can be made to work as a way of tackling the issue of adversariality. This is an important issue in argumentation, and perhaps an especially sharply focused issue in virtue argumentation, because there are a number of people who say that the adversariality of arguments is a problem. This stance is perhaps most familiar as a feminist critique but is not exclusively so (Rooney, 2010; Hundleby, 2013; Steenhagen, 2016). Such critics maintain that the way in which arguments tend to be couched, perhaps especially in philosophy, but in argumentation more generally, is needlessly damaging, because it is framed in adversarial terms. Trudy Govier has a response in terms of “minimum adversariality” (Govier, 1999, 55). Scott Aiken has refined his own version of this position over several iterations; here is a recent version:

4 At

least if “⇒” is understood as encompassing defeasible as well as deductive inference. I follow Dutilh Novaes in broadening MacFarlane’s definition in this manner (Dutilh Novaes, 2015, 590).

104

A. Aberdein

an arguer taking a critical line with one’s own commitments is, for the purposes of the argument, both an opponent and an ally. The thought is that without the role-related duties of critical dialogue, there are moves of critical probing that must be performed that are, in their dialectical function, oppositional. However, this is yet in the service of a broader cooperative goal of dialectical testing of reasons and acceptability (Aikin, 2017, 16).

In other words, if you are concerned that we keep framing people as opponents, remember that they are also allies. Here is a similar approach to the same idea from Dutilh Novaes, that she frames hopefully as “virtuous adversariality”, a strong cooperative, virtuous component which regulates and constrains the adversarial component. … The picture that emerges is of philosophy as ongoing conversation between interlocutors who respectfully disagree with each other. Disagreement forces them to provide the best possible arguments to support their respective positions, and it is from the comparison between the best possible arguments for competing positions that more robust philosophical theories emerge. Interestingly, Grice’s famous conversational maxims offer sensible principles on how to conduct such philosophical conversations, offering the right balance between adversariality and cooperation (Dutilh Novaes, 2014).

And, as a reminder, here are Grice’s famous conversational maxims: Quantity 1. Make your contribution as informative as is required (for the current purposes of the exchange). 2. Do not make your contribution more informative than is required. Quality Try to make your contribution one that is true. 1. Do not say what you believe to be false. 2. Do not say that for which you lack adequate evidence. Relation Be relevant. Manner Be perspicuous. 1. Avoid obscurity of expression. 2. Avoid ambiguity. 3. Be brief (avoid unnecessary prolixity). 4. Be orderly (Grice, 1975, 45 f.).

Most of these are covered by the assertion instance of Brogaard’s intellectual flourishing norm: You should assert p only if asserting p does not hinder intellectual flourishing. Indeed, Grice himself subordinates all of these maxims to a single norm: Cooperative Principle Make your conversational contribution such as is required, at the stage at which it occurs, by the accepted purpose or direction of the talk exchange in which you are engaged (Grice, 1975, 45).

It takes no great leap to see this as consistent with intellectual flourishing, at least as regards conversational exchanges. Certainly persistent and unmotivated disregard for this principle would seem likely to hinder intellectual flourishing.

7 Eudaimonistic Argumentation

105

7.4 Conclusion Most discussion of VTA has emphasized the “virtues in” side, the contention that we need virtues of a certain type in order to argue the right way. However, the case I have made here is that we should also hope for “virtues out”. Arguing well requires good, that is, virtuous arguers; but also arguing well, that is virtuously, is required for good arguers. If there are to be good arguers, there will need to be arguing well. More broadly, if intellectual flourishing cannot be attained in solitude, then there must be a community of interacting agents, whose potential for intellectual flourishing will only be realized individually if it is realized collectively. In any practical scenario, the interactions between these agents will exhibit disagreements and differences of opinion; that is, the basis for arguments. But these arguments ought to be fully satisfying ones, if the community is to continue to flourish. So, an indispensable component of any intellectually flourishing community will be fully satisfying arguments, which, if the treatment above is correct, will comply with instances of the intellectual flourishing norm. More broadly still, the satisfactory resolution of ethical debates will require such a community. Hence, if there are to be good people in general then there must be good argument.

References Aberdein, A. (2010). Virtue in argument. Argumentation, 24(2), 165–179. Aberdein, A. (2018). Virtuous norms for visual arguers. Argumentation, 32(1), 1–23. Aikin, S. F. (2017). Fallacy theory, the negativity problem, and minimal dialectical adversariality. Cogency, 9(1), 7–19. Bailin, S. (2014). Commentary on: Moira Howes’s “Does happiness increase the objectivity of arguers?”. In D. Mohammed & M. Lewi´nski (Eds.), Virtues of Argumentation: Proceedings of the 10th International Conference of the Ontario Society for the Study of Argumentation (OSSA), May 22–25, 2013. Windsor, ON: OSSA. Battaly, H. (2008). Virtue epistemology. Philosophy Compass, 3(4), 639–663. Brogaard, B. (2014a). Intellectual flourishing as the fundamental epistemic norm. In J. Turri & C. Littlejohn (Eds.), Epistemic norms: New essays on action, belief, and assertion (pp. 11–31). Oxford: Oxford University Press. Brogaard, B. (2014b). Towards a eudaimonistic virtue epistemology. In A. Fairweather (Ed.), Virtue epistemology naturalized: Bridges between virtue epistemology and philosophy of science (pp. 83–102). Cham: Springer. Brogaard, B. (2014c). Wide-scope requirements and the ethics of belief. In J. Matheson & R. Vitz (Eds.), The ethics of belief: Individual and social (pp. 130–145). Oxford: Oxford University Press. Clark, S. J. (2003). The character of persuasion. Ave Maria Law Review, 1(1), 61–79. Clark, S. J. (2011). What we make matter. Michigan Law Review, 109(6), 849–862. Clark, S. J. (2013). To teach and persuade. Pepperdine Law Review, 39(5), 1371–1399. Cohen, D. H. (2005). Arguments that backfire. In D. Hitchcock & D. Farr (Eds.), The uses of argument (pp. 58–65). Hamilton, ON: OSSA. Cohen, D. H. (2013). Virtue, in context. Informal Logic, 33(4), 471–485. Cohen, D. H., & Miller, G. (2016). What virtue argumentation theory misses: The case of compathetic argumentation. Topoi, 35(2), 451–460.

106

A. Aberdein

Dutilh Novaes, C. (2014). Virtuous adversariality as a model for philosophical inquiry. Presented at Edinburgh Women in Philosophy Group Spring Workshop on Philosophical Methodologies. Dutilh Novaes, C. (2015). A dialogical, multi-agent account of the normativity of logic. Dialectica, 69(4), 587–609. Govier, T. (1999). The philosophy of argument. Newport News, VA: Vale Press. Grice, P. (1975). Logic and conversation. In P. Cole & J. L. Morgan (Eds.), Syntax and semantics (Vol. 3, pp. 41–58). New York, NY: Academic Press. Howes, M. (2014). Does happiness increase the objectivity of arguers? In D. Mohammed & M. Lewi´nski (Eds.), Virtues of Argumentation: Proceedings of the 10th International Conference of the Ontario Society for the Study of Argumentation (OSSA), May 22–25, 2013. Windsor, ON: OSSA. Hundleby, C. (2013). Aggression, politeness, and abstract adversaries. Informal Logic, 33(2), 238–262. Hursthouse, R., & Pettigrove, G. (2018). Virtue ethics. In E. N. Zalta (Ed.), The Stanford encyclopedia of philosophy. Stanford, CA: C.S.L.I. MacFarlane, J. (2004). In what sense (if any) is logic normative for thought. Online at https:// johnmacfarlane.net/normativity_of_logic.pdf. Noddings, N. (1984). Caring: A feminine approach to ethics and moral education. Berkeley, CA: University of California Press. Norlock, K. J. (2014). Receptivity as a virtue of (practitioners of) argumentation. In D. Mohammed & M. Lewi´nski (Eds.), Virtues of Argumentation: Proceedings of the 10th International Conference of the Ontario Society for the Study of Argumentation (OSSA), May 22–25, 2013. Windsor, ON: OSSA. Rooney, P. (2010). Philosophy, adversarial argumentation, and embattled reason. Informal Logic, 30(3), 203–234. Steenhagen, M. (2016). Against adversarial discussion. Collingwood and British Idealism Studies, 22(1), 87–112.

Chapter 8

Worries About the Prospects for Community Argument Dale Hample

8.1 Introduction Many people believe that arguing is fundamentally competitive, perhaps even a verbal enactment of combat (Hample, Han, & Payne, 2010). A more sophisticated view, shared by many scholars, is that interpersonal argumentation is essentially cooperative. In arguing, people should share frames of reference, exchange turns at talk, and reply responsively to what the other person says (e.g., Brockriede, 1975; Gilbert, 1997; Jackson, 2017). Very often, people do these things successfully, or at least approach those norms. Such cooperative work is done in the service of seeking collaborative decisions, mutual choices about what to believe, feel, or do (van Eemeren & Grootendorst, 2004). Disagreement is often competitive, but it should take place on a platform of shared norms and assumptions. Certainly, arguments can derail and go bad, but this paper is about the prospects of them going well. A way of expressing the cooperative ideal for face-to-face arguing is to say that it should be community argument. Such arguing takes place within a mutual community of shared values and mutual respect, or invites an external person to conjoin the arguers’ communities during the argument. Johnstone (1982) called this second possibility bilateral arguing. This produces the mutuality of concerns and respect. Arguing’s great merits occur when it is used to regulate or explore disagreement. Disagreement is normally initiated by contact with the outside, with something alien to one’s current thinking. A pleasant rehearsal of beliefs runs into an inconvenient example, or a conversational partner surprises one with an expression of doubt. Disagreement and controversy are always new in this sense, and come from friction within one’s interior or social community. Community argument might be a comfortable conversation with agreeable intimates, but it might also be a social practice D. Hample (B) Department of Communication, University of Maryland, College Park, USA e-mail: [email protected] © Springer Nature Switzerland AG 2020 F. H. van Eemeren and B. Garssen (eds.), From Argument Schemes to Argumentative Relations in the Wild, Argumentation Library 35, https://doi.org/10.1007/978-3-030-28367-4_8

107

108

D. Hample

in which an outsider is invited to join and reinforce one’s community. At its best, arguing displays and creates community. Humans have evolved to participate in homogenous social groups, and the need for cognitive efficiency has generated evolutionary pressure to search only local memory to evaluate our views and decisions. The first is a bias against outsiders’ views and the other is a bias against thorough self-interrogation. These barriers to community argument are natural. They can be resisted, but only with considerable personal or social effort (e.g., public education). Other obstacles to community argument are more recent. These are emerging in social science research that is designed to prevent critical thinking about controversial topics. In the last half century researchers have been developing persuasion technologies that intentionally prevent people from thinking hard about persuasive messages. These soft technologies include distraction, narrative argument, visual argument, heuristics, nudging, and immersive virtual reality. They are all aimed at preventing counterarguing, an activity that is required for argumentation and critical evaluation of a viewpoint. Argumentation scholars are uniquely qualified to invent critical thinking principles that can be applied to all these persuasive technologies, not just to the linear verbal forms that have been our historical preoccupation. We should work on a generalized argumentation pedagogy that can be simply taught, so that people learn how to test arguments that would otherwise evade the possibility of thoughtful community. Those principles need to be equally applicable to syllogisms as to stories, to reasoning as to temptation, and to open expression as to hidden agendas. In this paper, I will first review the well-understood case for having community arguments. Then I will explore both the inherent and emerging obstacles to it. I will conclude by suggesting how our own community should try to address the problems I describe.

8.2 The Importance of Community Argument I sometimes use “community” as a metaphor here. The metaphor will refer to the movement of thoughts from one private memory store into another one. When I intend a literal usage, I mean to indicate that various people live together and connect freely with one another, so that when they exchange reasons they experience a community argument, aimed at community agreement. The fact that the people are part of the same community is not a small thing, and we will soon see how important it is. In the modern age, literal neighborhoods are shared by different sorts of people. They have different backgrounds and life histories, and often diverge in their beliefs, values, and judgments. When they talk or exchange reasoning with one another, they speak from different stances and reflect different views. They can be invited or taught to coalesce (Gilbert, 1997). These differences can supply special value to disagreement. We can review an argument that draws on nothing new. This might happen to a scholar who reads an

8 Worries About the Prospects for Community Argument

109

article that he or she wrote years ago. That sort of argument review is no longer inventive. It could remind us of something, but it does not really impel any new thoughts. However if we talk with an outsider—a person strange in his value judgments, odd in her life experiences, peculiar in his attitudes, mysterious in her allusions, and surprising in his reasons—that is when we are challenged. We may turn stubbornly away from the challenge, and then we will gain nothing because we risked nothing (Brockriede, 1975). But we might engage our neighbor in her suddenly apparent divergence, and it is then that we can improve our choices. Several things are required for satisfying community arguments. First, there must be difference, difference that leads to disagreement in the moment. In addition to this friction, we must also have mutual engagement, variously called clash or responsiveness (Jackson, 2017). Both arguers must have equal right to speak and equal access to key resources (Burleson & Kline, 1979; van Eemeren, Grootendorst, Jackson, & Jacobs, 1993, Chap. 2; Wenzel, 1980). And of course, we need to have enough in common for communication to take place (Brockriede, 1975; Sprain & Black, 2018). These are the conditions that have led to some of the great debates in human history: between democracy and the divine right of kings (Habermas, 1989), between Protestants and Catholics (Edwards, 2004), between subordinate women and dominating men (Beauvoir, 1972), between the sciences and the humanities (Snow, 1963), between private property and public interest (Carson, 1962), and a multitude of other great arguments that shaped our national and personal histories. We can also understand our own minds as communities. We have chunks of memory segregated from one another to varying degrees. Sometimes we pass easily from one community of knowledge to another, as when we think about automobiles and then about mass transit. Sometimes it is more effortful to move from one to another, as when we think about automobiles and then Roman aqueducts. I will also use “community” metaphorically to include this interior passage from one portion of our mind to another because for the purposes of this paper, difficult mental connections are enough like moving between one’s own thoughts to a neighbor’s that I can make similar points about both processes. Arguments without controversy can supply entertainment, self-esteem, or social identity (Cho, Ahmed, Keum, Choi, & Lee, 2018; Hample, 2018, Chap. 2). But they rarely generate intellectual growth or change. Whether we are interested in personal or social development, community argument is necessary. Divergence and reasoning must join together with the aim of one changing the other. They are both essential for the development of new meanings, which is exactly the province of serious argument (Hample, 2009). Thus, I am convinced that community argument is a necessary element of civilization, whether national or individual.

8.3 Inherent Barriers to Community Argument This paper expresses my worries about the possibility of community arguments. My concerns are mainly in two classes: those that originate in our evolutionary heritage,

110

D. Hample

and those that we are actively creating in our current scholarship. Both produce threats to the possibility of vigorous, divergent, reasoned disagreement taking place in the context of mutual respect. In this section I will examine two inherent problems for community argument, obstacles that human evolution has insisted upon. Those worries concern our need to live together, and the compromises we must make in order to think quickly.

8.3.1 Evolved Sociality Humans evolved as a social species. Both our primate ancestors and early versions of our own species lived in small tribes or family groups. Fagan (2010, p. 6) suggested that a typical Neanderthal never came into contact with more than a few dozen other humans in a lifetime, and that most of the out-group contact was very brief (also see Simpson & Belsky, 2016, p. 94). Often it involved raids or trade. A person’s life was tightly composed in regard to social relations. Every individual depended on others for hunting, for gathering, for exploring, for sheltering, for mating, and so forth. Today we think of our social groups as sources of interpersonal stimulation and emotional pleasure, but in our environment of evolutionary adaptedness the kinship group was a decisive determinant of life or death. People who lived comfortably in groups were favored evolutionarily. Loners, people who were uncooperative, those who would not do their part—these people either left their tribes or were shunned (Cosmides & Tooby, 1992; Roos, Gelfand, Dau, & Lun, 2015). Alone, one’s survival prospects were slim, and life expectancy would be measured in days or weeks. Survival and reproduction required cooperation, conformity, reliability, reciprocity, and trust—all within the group (Bateson, 1988). But this cannot be the whole story. As we contemplate our own lives, we see that argument is all around us. Disagreement, withholding cooperation, not taking the other at his or her word, insisting on a different decision—how could all of these also have developed as psychological adaptations in the face of these fundamental needs for sociality? Mercier and Sperber (2017) developed an argumentative theory of reasoning. Many people easily suppose that humans first learned to reason and then learned to say it out loud. But Mercier and Sperber proposed the reverse: that arguing originated in response to persuasion. Speaking first, thinking second. The simplest sort of persuasion is a straightforward command or request, a more nuanced version includes a reason for the desired action, and a much more sophisticated kind of persuasion provides a good reason. Mercier and Sperber said that people developed epistemic vigilance so that they could distinguish these sorts of communicative events from one another, and thus to protect themselves against unreasonable exploitation. As millennia passed, persuaders learned to anticipate critical reactions and thus began to argue better. So from an evolutionary point of view, arguing has always lived in the friction between two powerful energies: sociality required cooperation and agreement, and self-defense required reasoning and reserve. The inevitability of difference among

8 Worries About the Prospects for Community Argument

111

humans collided with the need to merge oneself into the tribe. Rejecting the kinship group would be fatal, and inability to defend oneself against self-interested persuaders would be fatal as well. People who could not maneuver between these pressures did not become our ancestors. Of the two forces, sociality is the one that presents the evolved obstacle to community argument. Spoor and Kelly (2004) said that our early tribal communities were oriented to group affect, and that positive feelings were one of the factors that kept people together. A key sort of affect was cohesiveness, which had three elements: interpersonal attraction, commitment to a task, and a sense of pride in one’s group membership. Here again we see the idea of cooperation and the impulse for agreement as a foundation for group life. To disagree, to reject, to doubt, to refuse— these are all contraindicated by cohesion. To bond is to agree, and so agreement is pleasurable. This sociable wiring continues to affect human life today. A well-established result in social psychology is that people have a “need to belong,” which has the evolutionary basis just reviewed (Baumeister & Leary, 1995, esp. p. 499). Today, isolation is connected to poor physical and mental health, and people are distressed at the prospect of a social bond breaking (e.g., by divorce or geographical separation). A meta-analysis by Holt-Lunstad, Smith, Baker, Harris, and Stephenson (2015) showed that social isolation increased the likelihood of mortality by about 30%, even in our modern age. Evolution established the norms and essential behaviors for interpersonal connection from birth to old age, and so continues to outline the organization of our families and close relationships (Simpson & Belsky, 2016). Our social lives can be partly described in terms of our continuing patterns of attachment impulses, and these are strongly affected by our trust and value for others. Our attachment behaviors mutually orient us in each of our relationships. And more importantly, these motivations cause us to seek affiliation and to be disturbed or hurt by a lack of attachment. Belongingness, a highly sought goal of sociality, reverberates though the human condition, and we can see how it influences arguing. The need to belong can spontaneously transform into a need to step back and pacify because human social systems had clear hierarchies of power and status (Boehm, 1999). Power distance, the willing acceptance of those social disparities, continues to affect people’s willingness to argue today (e.g., Santabañez, Hample, & Hample, 2019). When people are faced with danger, argumentation scholars might recommend that one try to manage the threat with arguments, but this is a minor suggestion in view of the main impulses evolution has provided: fight/flight or tend-and-befriend (Taylor, 2012). Fighting might be argumentative but might also be physical or otherwise unreasoned. Flight is avoidance, and so rules out any sort of interaction. Tend-and-befriend is the impulse to rush into one’s group in response to stress or threat. Very generally speaking, evolution has taught men to fight or flee and has taught women to retreat to a condition of social support that throws up a barrier to invaders (Taylor, 2012). Tend-and-befriend is a particular sort of motivated avoidance, in which one shelters from an outside challenge—physical or argumentative—and seeks protection within one’s most valuable

112

D. Hample

social group. Tend-and-befriend, flight, and maybe fight have in common that they shun argumentative involvement. We see all these impulses still strongly at work today. People prefer to interact with people they agree with and therefore people who can be depended upon to agree in return. Normally we seek to bond, not to debate. People are threatened by the prospect of discussing divisive issues (Simons & Green, 2018). Noelle-Neumann’s (1974) “spiral of silence” theory said that the fear of social isolation would make people reluctant to express unpopular opinions. A recent meta-analysis (Matthes, Knoll & von Sikorski, 2018) has shown that the spiral of silence is reliably apparent in human behavior and intentions. The modest overall effect (r = 0.10) was dramatically increased (r = 0.34) when people were confronted with the possibility of expressing a minority opinion about a personally important matter to one’s friends or family. The more important the social bond, the more attractive was the silence. The ideal of community argument, however, objects to silence and requires that people come into serious contact with others who disagree. In our oldest evolutionary environments, this might amount to interacting with a member of another tribe. Here again, we see innate pressures not to engage productively. Mackie, Devos, and Smith (2000) explored the connections between group identification and reactions to an outgroup. Respondents were asked to identify themselves as members of the “group” that had particular views—e.g., that drug offenses should be strongly punished—but received no argumentation from the other side. Just knowing that the other “group” disagreed with them produced significant feelings of anger and contempt toward that group, as well as intentions to confront or attack the other group. These effects were particularly pronounced when respondents felt that their in-group was in a strong position regarding the issue at hand. This reaction, of course, systematically discourages productive arguing across groups and makes an invitation into one’s own community less likely. Those group affiliations were artificial, of course. Political parties are more natural tribes today. Gervais (2015) exposed Republicans and Democrats to online comments that were manipulated to be civil or uncivil, but all critical, regarding an Obama Administration policy. The Democrats became angrier than the Republicans when exposed to the nastiness, and were also more likely to critique the author of the Obama-opposed comment. This is consistent with the idea that the Democrats were emotionally protecting their home group (the Obama administration), but the Republicans were not. This has immediate implications for whether any argumentation is likely to be productive. Molina and Jennings (2018) tested the effects of civil and uncivil online remarks on people’s willingness to engage in the deliberation. They found that civil comments led to greater mental elaboration of the issues, and then to more willingness to engage. But we see here that outgroup incivility led to anger and reactivity, and repressed high quality argumentation with challengers. Arguing, especially arguing with an outsider, simultaneously confronts two basic human impulses: the need to be sociable and the need to protect, to seek mutuality and to insist on selfish advantage. Wrangham (2019) showed that humans are commonly tolerant and peaceful within a clan but murderously aggressive toward outsiders. One might suppose that within-tribe violence might be an evolutionary advantage: a

8 Worries About the Prospects for Community Argument

113

strong male could kill another man to take his wife, or beat a woman and take her food. Violence and competition are part of our biological heritage, shared with various of our primate cousins. Why aren’t we all killers? Wrangham argued that our species is self-domesticated: we turn our violence outward but guard against it at home. In fact, when someone was too aggressive, too violent, too dangerous within the clan, the other men would join together to kill him. In contrast, warriors who ruthlessly protected the group against external competition were favored. Thus clans regulated their own genetic estates by insisting on the right balance of local agreement and external aggression. The analogy between violence and argument is unfortunately not a distant one, since we so easily describe interpersonal disagreement in military terms (Jensen, 1983). We want to protect the agreeability of our home, and guard it against outsiders. These impulses lead us both to avoid disagreement with our familiars and to resist genuine engagement with those who might threaten our consensus. So this cursory summary of sociality, one element of our evolutionary heritage, shows that we evolved to need reliable social bonds, and that those require trust, agreement, and the attractiveness of joint action. Our very survival depended on our motivation and ability to maintain membership in close tribal groups. We continue to act that way today. We try to find groups for ourselves, communities of likeminded people with whom we are comfortable in a literally agreeable relationship. We find disagreement threatening and build barriers against it. In our chosen daily lives, we prefer cooperation to confrontation, and are hurt and disturbed when our groups begin to flounder or dissolve. Nastiness is a natural response to a disagreeable out-group member, but only civility invites good arguing. We have evolved to meet contradiction with silence. Little of this makes us seem eager to engage in community arguing. Biology is not destiny, and the inevitability of human differences means that there is also a countervailing pressure that favors arguing. But we should recognize that passivity, avoidance, and insincere agreement are evolved adaptations that are in service to our larger social lives. These things can be addressed with large-scale interventions, such as K-12 public education, but superficial complaints about people occupying “echo chambers” are no more than shouting into a hurricane.

8.3.2 Evolved Cognitive Modularity To this point I have been considering “community” in a fairly literal way. I have been thinking about Pleistocene clans and tribes, and contemporary families, neighborhoods, and identity groups. I have been contemplating the consequences of members of those communities not arguing with one another or their more distant neighbors in a constructive way. But part of Cherniak’s (1986) argument suggests that another sort of community, this one more metaphorical, also had important relevance to whether or not we argue well. These “communities” are the little sections of our minds that operate with easy modularity and only effortful cross-module coordination. I am

114

D. Hample

proposing that we think of each module as a community, and that we consider what happens when the communities do not easily speak to one another. A number of psychological theories deal with the “modularity of mind,” a theoretical orientation that began in modern times with Fodor (1983) and has much older roots in faculty psychology (e.g., Campbell, 1776/1988). The basic contemporary idea is that evolution required us to develop specialized information processing abilities so that people could avoid danger, detect cheaters, recognize safe habitats, remember words, form sentences, and so forth. Issues that directly bore on survival (and human bonding is part of that, as we have seen) needed their own quick, simple, and normally accurate solutions. The modules are distinct both biologically and functionally, so that verbal abilities have particular brain locales, as do musical abilities or spatial reasoning. We all understand how one module’s activity might not participate in a community way with that of another module. For instance, if a children’s song gave directions for finding your way home in a forest, people might well enjoy singing the song without absorbing anything about navigation in the woods. Kenrick (2012) said that for both humans and other animals, minds developed with different functions clearly separated so that they can operate without interference from other modules. He offered the example of birds that learn birdsong and what foods are poison in starkly different ways, and that both of these learning procedures are unrelated to how birds remember where good food is that day. This is in the service of speed and efficiency, so that conclusions that were urgent in the Pleistocene’s environment of evolutionary adaptedness would appear as immediately and reliably as possible. Here I want to concentrate on a special modularity, the one that Cherniak (1986) discusses at length—the modularity of long-term memory (LTM) stores. Humans have two kinds of memory, short-term (STM) and long-term memory. Our active thinking goes on in STM, also often called working memory. We can bring about half a dozen things into STM at once, and we shake them together, making inferences, noticing consistencies, realizing how one is an exception to another, and so forth. All these things (even the names of percepts) have to be retrieved from the long-term memory inventories. Essentially, our thinking goes on in working memory, which is therefore the site for producing, evaluating, and comprehending arguments. LTM, on the other hand, is like a warehouse in which innumerably more than half a dozen things are stored. The items on the shelves of long-term memory are essentially inert, dissolving slowly unless the forklift comes around periodically to take them into the STM showroom. If all the relevant items are not retrieved from LTM, this means that some memories have no voice and so are uninvolved in our immediate thinking. When that happens, we can have something analogous to a failure of community interaction. The design problem is that the forklift needs to find things efficiently. LTM searches cannot be done randomly because that would be hopelessly inefficient, and our ancestors would all have died thinking about ponds or tubers because the forklift failed to return anything involving predators that live in caves. LTM has to be organized somehow, just as we put particular files in particular folders on our hard drives. Most of us look for files by going to a particular directory, looking at

8 Worries About the Prospects for Community Argument

115

the list of subfolders, and then scrutinizing the filenames in the list. Hard drives, libraries, academic conferences, and our own memories—all of them have labeling, indexing, hierarchy, and networking, and all of this is in the interest of efficient retrieval. Random lists are the opposite of organized materials, and using them is inefficient, tedious, and often inconclusive. In the early days of personal computers, some people stored all their files in the root directory, but we soon found that our work smoothed out when we learned to create directories and subdirectories, using names that informatively indexed the contents. Some files—some chunks of memory—are in the same folder or mental community and some are in remote ones. The further the forklift needs to travel, the more time it takes, and the decision may not wait until the search is over. The point of organizing, remember, is that we do not have to search the whole warehouse. We find what we need and quit looking. With good organization this will be productive and with bad organization we may walk into a predator’s cave. This is part of the reason that Cherniak’s (1986) theory is called minimal rationality (for other explicit application of Cherniak’s theory to argumentation, see Canary, Brossmann, Brossmann, & Weger, 1995; Weger & Canary, 2010), which is a descendant of Simon’s idea of bounded rationality (Crowther-Heyck, 2005). There would be no point in bothering to organize memory if we continued to search once we found what we thought we needed. But neither the memory organization nor the search strategies are guaranteed to be perfect. Evolution is oriented to the species’ welfare, not to the individual’s. Cherniak (1986, p. 50) made this point clearly: “…part of the human condition is in fact to fail to ‘make the connections’ sometimes in a web of interconnected beliefs, to fall short of a synoptic view of one’s belief system.” That is, human memory is not one liquid volume or force field in which any perturbation (a perception, someone else’s excellent argument) automatically ripples out to every other relevant element. The idea that we all have a unified “web of belief” is regrettably persistent in our philosophical reflections about memory (Gilbert, 2014, p. 84; Quine & Ullian, 1978). Short- and long-term memory are distinct subsystems, and LTM itself has an unmistakable organization in which various chunks are walled away from others. Only material in STM can be “worked on:” the contents of LTM are dormant and out of consciousness. Thus if an arguer suddenly realizes that the other person is right, only material that is already in STM (or that is quickly retrieved) will be altered, and the alteration has to persist through the process of returning the memory to the long term inventories. Sometimes a better thought might be properly acknowledged in working memory without necessarily being stored, as many parents and spouses can attest. Strength of association predicts ease and likelihood of the materials being connected to one another, and thus able to inform one another. Associations are stronger within chunks than between them. Cherniak (1986, p. 57) related the example of a man who understands that a match will ignite gasoline but nonetheless uses a lit match to peer into a gas tank to see if it is empty. Cherniak’s explanation was that “means of ignition” was in a folder not closely associated with “means of illumination.” As time and life experience accumulate, “…the inactivity of the beliefs in long-term memory in itself degrades rationality. It results in the accumulation of unrecognized

116

D. Hample

inconsistencies, valuable inferences not being made, and so on. Only the behavior of the contents of short-term memory can counterbalance the results of the inertness of the beliefs in long-term memory and so contribute to the maintenance of adequate rationality” (Cherniak, 1986, pp. 60-61). So to state my point: We have evolved a mental organization that serves us efficiently but not with perfect rationality. Whether arguing alone or interpersonally, we may forget to mention some important things, we may not realize that we are saying something that we know to be wrong in another context, and we may, in the moment, not bring to bear enough of our own experience to test another person’s views properly. These circumstances are analogous to those we noticed when we realized that we prefer to agree with people who are like us, that we exaggerate the badness of out-group arguments, and that we do not seek out people who can give our tribe’s arguments their most demanding tests. These are not thinking flaws that derive from the narcissism of millennials, the superficiality of youth, the profit focus of modern technology corporations, or the closed-mindedness of the other political party. These are inherent remnants of our evolutionary heritage. We can rise above them, but only with very substantial interventions that take seriously why we have these apparent flaws in critical thinking.

8.4 Emerging Barriers to Community Argument Here I change my chronological focus from the Pleistocene Era to the present day, and I turn my attention away from the unprejudiced workings of evolution to the intentional goals of a large group of social scientists. These researchers do not seem to have noticed that their work has ethical implications. They are building soft technologies of persuasion without wondering enough what the machines will be used for. The fundamental problem, as I see it, is that considerable research now targets counterarguing as a phenomenon to be sabotaged. Suppressing counterarguing—the critical testing of other people’s persuasive messages—fundamentally opposes the possibility of vibrant community argument. I will review research from the last half century, research that has been aimed at the development of persuasion strategies that are intended to eliminate or reduce critical thinking on the part of the persuasion’s audience. Having reviewed the importance of community arguing in the first portion of the paper I think I do not need to establish here that anything aimed at corroding critical thought is ethically questionable.

8.4.1 Distraction As far as I can tell, this history begins in the 1960s with some early research on distraction. In a classic study, Festinger and Maccoby (1964) provided a persuasive message to two experimental groups. One group was distracted from the message

8 Worries About the Prospects for Community Argument

117

by a silent movie that was being played concurrently. The other group received the persuasion under more ordinary conditions, with no experimental distractions. The distracted group was more persuaded. Festinger and Maccoby explained this outcome by suggesting that the distracted group had been unable to counterargue. They paid just enough attention to the message to have been reached by it, but not enough attention to think about it properly. Their published remarks might well have been the first scholarly notice that messages can be more effective when we can remove or reduce the capacity of an audience to think about the message’s content. The connection of this distraction effect to counterarguing was substantially sharpened about ten years later by Petty, Wells, and Brock (1976). They pointed out that distraction’s suppression of counterarguing should promote more persuasion if the counterarguing would have resisted the message’s point, but that the suppression might also lessen persuasion if the people would otherwise have had positive thoughts about the message. Their pair of experiments supported this theory. They concluded that “distraction is most likely to lead to enhanced persuasion when a message presents poor arguments (i.e., arguments that are open to refutation and counterargumentation) and to reduced persuasion when a message presents very good arguments” (p. 883). Thus distraction shields bad arguments from a community’s argumentative energy. The distraction research now continues under a more contemporary label, one that might invite less obvious objection from those of us interested in the quality of public discourse. Jeong and Hwang (2012) studied “multitasking” of the sort that students do while studying. Their experimental groups experienced one of three conditions: no multitasking, multitasking with a primary focus on the persuasive message, and multitasking with a secondary focus on the message. The written messages were constructed on the Toulmin model so that a reasonably capable person could detect flaws in either the data or warrant of the arguments. The experimenters used a portion of a Transformers movie as the “other” task. Results showed that the non-multitasking group showed far more counterarguing than the distracted conditions (η2 = 0.12 or 0.20, depending on the operationalizations). Students distracted by the movie were less able to defend themselves against the poor arguments in the persuasive message. While Buller’s (1986) meta-analysis of distraction effects revealed a somewhat choppy record of designs and findings regarding counterarguing, he reported an overall correlation of r = −0.17 between distraction and counterarguing, and a correlation of r = −0.50 between counterarguing and attitude change. So the more distraction, the less counterarguing, and the less counterarguing, the more persuasion. The first result bears only on distraction as our first documented impediment to counterarguing. The second result, however, shows how strongly one can increase adherence if only one can reduce critical thinking. A value of r = −0.50 is quite substantial for social science research, and is a stronger effect than one usually uncovers in persuasion research. In Shen and Bigsby’s (2013) survey of the persuasive effectiveness of many message features, not a single one of the meta-analytic effects they report is as high as 0.50. So half a century of research has found the first persuasive tactic I am considering here. That is distraction (or multitasking). Argumentative content might be somehow required, perhaps by an advertising regulator or by public expectations about how a

118

D. Hample

political campaign is to be conducted. But that content might be paired with startling images, bouncy music, an exciting video, or a celebrity. This is a systematic way of suppressing counterarguing, and therefore a tactic for making weak persuasive messages more effective. I have not discovered any papers on this topic where the authors worried about the ethics of distracting audiences so that they thought less critically.

8.4.2 Nudges Distraction erases critical thinking by the simple expedient of getting people not to pay attention well enough to develop thoughtful reactions to a persuasive argument. Let me continue that theme by considering nudging. This idea generated a lot of public discussion with the publication of Thaler and Sunstein (1999), but many teachers of communication and social psychology had made contact with the basic idea by reading or teaching from Cialdini’s (1984) book on persuasion tricks, a book that has now moved through many editions, revised titles, and newer publication dates. Cialdini’s tricks involve few arguments, and none at the substantial center of the decision. He showed how people’s behaviors could be changed by giving them little favors, getting them to agree to something not as weighty as what is wanted, describing a product as scarce, and so forth. He did not discuss the tactic of giving good arguments to favor a stance, and in fact generated a set of tactics that detoured around the possible need to justify what one says. The basic idea of nudging is that people can be persuaded toward some behavior by means of “choice architecture,” which is essentially some design feature that encourages the behavior without attracting thought or using explicit argumentation. A prized example is that organ donation choices became enormously more common in France when citizens’ default choice changed from “don’t donate” to “donate.” The government claimed the “automatic” ground in this decision without needing to make an explicit argument to those being asked to donate their organs. Johnson et al. (2012) summarized the nudge idea as having two types of strategy: structuring the choice or structuring the options. Structuring the choice involves things such as deciding how many alternatives to offer, supplying carefully chosen online information aids, or setting defaults. Presenting the options to be chosen among may involve placing them into strategically described groups or making certain attributes foregrounded or backgrounded. Notice that none of this involves arguments for or against any of the decisions. The persuader merely arranges things in a “suitable” fashion and walks away without needing to say anything that a person might think critically about. The persuasion works by arrangement, not argument. Lest readers be skeptical about my worries that the persuasion research I am summarizing will be implemented by professional persuaders such as political or corporate interests, be informed that these national governments proudly make official use of nudging for what the administrations feel to be our public interests: the U.S.,

8 Worries About the Prospects for Community Argument

119

Australia, the U.K., the Netherlands, Germany, India, Indonesia, Peru, and Singapore (BETA, 2019; Grunwald, 2009; Halpern, 2015; Social and Behavioral Sciences Team, 2015; Stroeker, 2016). International institutions such as the U.N., the World Bank, and others also nudge. More than a few commentators have noticed the ethical implications of persuading people by arranging rather than arguing (e.g., Campbell, 2017). My argument in this paper does not depend on my personal balance of skepticism versus cynicism about world governments, marketers, and high-revenue non-profit organizations. I will say that my confidence in political parties and their progeny is not so high that I am comfortable with every national administration using nudging to pursue its own immediate ends. Whether you agree with my hesitations or not, however, I merely want to point out that since the key element of nudging is arrangement, the persuader is not providing anything that can easily be counterargued. Re-arranging a list of options is not at all like providing evidence, and neither is specifying which choices require citizens to go to some level of trouble to opt out. Critical thinking is suppressed, on purpose.

8.4.3 Narratives Stories have points to them—perhaps a moral point, perhaps an insight about human nature, or perhaps a behavioral recommendation. Stories can be told so that they feature explicit arguments favoring their conclusion, but more often they simply perform or display the argument’s point (cf. Lake, 1983; Olmos, 2013). When the narrative merely enacts the desired behavior, we can see that this is at least a cousin to nudging, because the point is made by arranging the reader’s imagined experience, not by making one’s reasoning apparent. To model a behavior is not necessarily to argue for it. Bilandzic and Busselle (2013, pp. 200–201) were briefly uncomfortable with this contrast between narrative and linear argument, but resolved their worries by pointing out the dichotomy was untenable because arguments could include narrative examples and full-scale stories might contain arguments. They did notice that people are not always aware that a narrative might have a persuasive aim, and that this was likely to suppress counterarguing. They concluded the relevant section of their review by saying, “…the suppression of counterarguing, facilitated by a text’s narrative properties and a reader’s narrative engagement, has been discussed as the most important mechanism for narrative persuasion” (p. 208). Since no linear arguments may be obvious or even present in a narrative, we should not be surprised that Cin, Zanna, and Fong (2004) indicated that one of the sources of narrative’s persuasive effectiveness is that very little counterarguing takes place. Narratives can be conveyed in novels, short stories, poems, ballads, and simple self-disclosures, but quite a bit of the academic interest in the past few decades has concerned what is called edutainment. Governments or other influence agents influence the plot lines of televised dramas or similar vehicles to display safe sex practices, how to deal with school bullies, what to do when your spouse opposes

120

D. Hample

vaccination for your child, and similar topics. Shen and Han’s (2014) meta-analysis showed that edutainment modestly promotes persuasion in health contexts (r = 0.12, k = 22, n = 19,517). Slater (1997; Slater & Rouner, 2002; also Moyer-Gusé, 2008) made a fundamental theoretical connection between narrative persuasion and the elaboration likelihood model of persuasion (ELM). The ELM proposes that people are persuaded (or not) by careful scrutiny of a message’s arguments when they are engaged by the message’s topic, but are influenced (or not) by peripheral elements of the message (e.g., source credibility or attractiveness) when they are less engaged. Slater says that when people become caught up in a story or entranced by a character in it (i.e., they are “transported”), they cease to scrutinize the message—either its narrative development or its point. Thus, good narratives reduce counterarguing. The empirical record for this prediction is not clear-cut, but perhaps we should not have expected it to be. The degree to which counterarguing is suppressed is affected by the narrative’s power to transport, as well as by the degree to which people can detect that someone is trying to persuade them. All of the empirical studies use only a few narratives, and we know this will prevent secure generalizations. And most fundamentally, it will never be clear what the narrative is being compared to because narratives inherently cannot contain precisely the same information as a linear message aimed at the same point. Bilandzic and Busselle (2013) pointed out that counterarguing has been measured in quite different ways in this research tradition, and that fact has also contributed to the inconsistency of results. Let me offer a sample of the empirical research on the relation between narratives and counterarguing. Niederdeppe, Shapiro, and Porticella (2011) compared three written messages about obesity: a narrative, an evidence-based argument, and a combination of the two. They found that the narrative condition generated fewer counterarguments than the evidence condition, but only for people who self-identified as liberals. The counterarguing suppression only appeared for a certain group in that study. It did not appear at all in another investigation. Moyer-Gusé and Nabi (2010) showed videos about unplanned teen pregnancy to teenagers in the U.S. One of the videos had a narrative structure and the other did not. They found that being transported by the narrative actually increased counterarguing, contrary to predictions. They also found that counterarguing was strongly predicted by viewers’ levels of psychological reactance, implying that narratives might successfully be targeted to people who have little disposition to be reactive. However, we should notice that counterarguing is only one form of message elaboration. Elaboration can also be supportive of a message’s point (Petty et al., 1976), and we should be just as interested in pro-arguing as counterarguing. Shen and Seung (2018) discovered that self-reported message elaboration was very strongly predicted by transportation and related measures in their study of four videos dealing with sexual health. Empirical research does not indicate that edutainment always or strongly suppresses counterarguing. I suspect that the common conclusion that it does (e.g., Cin et al., 2004) may be more influenced by the attractiveness of the original theoretical reasoning than by the empirical record. In addition, the tests of narrative’s persuasive

8 Worries About the Prospects for Community Argument

121

power are hard to cumulate because the research designs have rarely been comparable. Bilandzic and Busselle (2013, pp. 202–203) found no empirical pattern of results consistent enough to justify a clear conclusion about whether narrative persuasion worked better than various alternatives, such as use of statistical evidence. In fact, it is hard to imagine constructing two messages with identical content but varying only in form (linear vs. narrative), so we should probably not expect any clear-cut findings about the merits of narrative framing. It is evident, however, that narratives can persuade and that sometimes they do so without the audience being aware that someone is trying to influence them. In spite of its unevenness, the research literature does seem to show that a suppression of counterarguing is expectable in some conditions, which of course may be exploitable by a persuader who can distract, nudge, or target audiences to magnify a narrative effect. The more involving narratives seem to have considerable potential for stimulating some sort of elaborated thinking by the audience. Since narratives only need to perform a point rather than argue for it explicitly, institutional persuaders may not find it hard to control what sorts of thoughts are invited. Whether social scientists succeed or not, these researchers are trying to find ways to use narrative to generate more persuasion by inhibiting counterarguing. In the case of narrative persuasion, we have a long-standing theory of argumentative analysis, Fisher’s (1987) suggestion that we focus on both the internal consistency of a story (its coherence) and the veridicality of its reference to the external world (fidelity). I think this pair of considerations has promise for critical analysis of narratives and perhaps other sorts of persuasion that suppress critical thinking. Narratives mainly show, they do not reason. But Fisher’s system can be used to force them into the mold of appeals that can be reasoned about.

8.4.4 Visual Persuasion Some forms of narrative such as edutainment combine a narrative schema with a visual presentation. For some time, argumentation scholars debated whether one could reasonably categorize a visual production as an argument. Today, I believe the debate is over and the consensus is that photographs, cartoons, films, and other images are (or at least can be) arguments (e.g., Birdsell & Groarke, 1996; Groarke, 1996). The reason for the initial debate was that argumentation theorists remained committed to their own educations, which presupposed that an argument was a linear verbal construction that featured explicit premises and conclusions. The argumentation community was able to work through enthymemes and warrants reasonably well, but got stuck for a while when we were confronted with what appeared to be entire arguments that had no verbal expressions or any proof claim at all. Once we were able to understand how to find the “point” of a drawing or photo, we got to work and learned to move backwards from the visual claim. Current research takes visual-as-argument for granted, and generally tries to work out how to analyze and critique such arguments (e.g., Dove, 2016; Godden, 2013; Ripley, 2008).

122

D. Hample

But for my present focus on counterarguing, I want to revisit why it was originally hard to reconcile visual art to argumentation theory. The basic complaint was that visual material wasn’t verbal—that premises were hopelessly camouflaged, inferences seemed magical, and conclusions could not be securely tracked back to anything. In communication theory this is sometimes expressed as a contrast between digital and analog information (Watzlawick, Beavin, & Jackson, 1967). Watzlawick and his colleagues explored distinctions between verbal material, which has digital meaning, and nonverbal behavior, which is analog. For instance, a digital computer (the kind most of us have) stores precise numbers, but an analog computer simply takes a continuous input from one device and passes it on to another. A traditional thermostat, for instance, is an analog device that takes continuous temperature input (registered by the bend in a sandwich of two strips of different metals) and passes it on to the furnace in the form of continuous adjustment instructions. Language is digital, and a proof of this is that words are to be found in dictionaries that list precise meanings for each word, even if there are many possible definitions. Nonverbal behavior is analog, and a proof of this is that nonverbal behaviors do not have dictionaries. For instance, a smile can have innumerable degrees. At one point it is alert; at the next point it is polite; at a further point it is friendly; and at yet another point it is welcoming. Those “points” cannot be found in a dictionary. They are more like the continuous information displayed on an old wind-up pocket watch than they are like a machine translator moving words from one language to another. Using this helpful distinction, we can see that propositional arguments are digital and visual ones are analog. With this background, consider a political cartoon. At what point does a politician’s nose become Jewish, and the cartoon become anti-Semitic? At what point is the politician’s waistline exaggerated so that he becomes greedy? At what point does the curve in the politician’s hands become dangerous, so that her extended gesture becomes a choking threat rather than an invitation? These are all interpretive problems and they all have the same nature. In every case I am asking that feathery analog information be translated into granite digital meaning. I can see what the cartoon is conveying, in just the same way that I can regard an internet meme as annoyingly cute. But I cannot make a sentence that precisely carries what I feel into what I can say. I can convey my intuition or my feeling (cf. Gilbert, 1997, 2014), but that is different from being able to write down a clearly examinable verbal argument. So if the visual argument is something like a heap of impressions, how does one counterargue against it? That is the crux of the matter for this paper. Just about any objection that is levied against a visual argument can be answered by replying, “I never said that” or “That’s only your take on it.” Any subtle point is deniable, and we are aware of that whenever we start to resist a poster, or a movie, or some protest graffiti. The extreme propaganda cartoons of decades and centuries past may now seem so extreme that they might as well be explicit propositions (see the many illustrations in Jowett & O’Donnell, 2019). The subtle drawings, the modestly retouched photos, and the carefully posed “casual” visual records—those are the things that need genuine thought, and the ones that most evade it. As I mentioned above, some members of our community are working out how to answer a visual argument, but I

8 Worries About the Prospects for Community Argument

123

suspect that the most typical reactions to a visual argument are either to accept it or to walk away. If I am right about that, the only possible outcomes for a visual persuader are to win or to draw. A visual argument cannot be lost, because engagement may seem pointless. A counterargument might find no grip on the deniable surface of a visual product. Even if a critic believes that he or she has discovered reasoning in a bit of political artwork, the argument can be disowned and denied. Counterarguing against such a message is like punching air in a dream. I am doubtful that people will naturally try.

8.4.5 Some New Horizons Communication technologies are now rapidly changing or being invented, and naturally these all carry affordances for persuasion. Anything that can convey messages can carry persuasive intentions to audiences. Many scholars are working on these new media and devices, and only a few are exploring their persuasive possibilities. But some do, and elements of the new research continue to see counterarguing as persuasion’s enemy, and suppression of thinking as a badge of persuasive effectiveness. In 2002, the U.S. Army began distributing a free video game called America’s Army. This was a first-person shooter game, and it effectively invited young men and women to immerse themselves in military adventures designed by the Pentagon. It was straightforwardly designed as a recruiting tool. Advertisers are now beginning to generate computer games for children to play (e.g., Waiguny, Nelson, & Terlutter, 2012), and have already found that the degree of children’s involvement in the games is important to whether the advertising works. This finding is similar to the narrative tradition’s claim that transportation was the key to suppression of counterarguing and persuasive success. Immersive virtual reality, using either elaborate equipment or newer smartphones, will surely be attracting more and more research attention in the years to come. Blascovich et al. (2002) have written a thorough theoretical review of how to integrate this technology with social psychology, and have identified “social presence” as a key process. Social presence belongs to the family that includes transportation, identification, and parasocial interaction, all of which are now theorized to have implications for counterarguing. Much of the research on these psychological processes was done to explore narrative persuasion and edutainment. Focused research evaluating the use of immersive virtual reality to persuade people has already begun, and has shown effects for the realism of digitized persuasive sources (e.g., Guadagno, Blascovich, Bailenson, & McCall, 2007). The theoretical commitments involved in this work appear to be cousins to those developed in the study of narratives, and so an interest in suppressing counterarguing will be natural. Already, one study has explicitly related immersive virtual reality to counterarguing. Two pairs of persuasive narrative messages were prepared, an immersive virtual reality version and an ordinary two-dimensional video. The paired messages only

124

D. Hample

differed in their “virtuality,” because one was a converted version of the other. People watched one or the other. For both topics, the immersive virtual reality version resulted in substantially less counterarguing by respondents (Ma, 2019). The experiences of computer gaming and immersive virtual reality are comparable to those of movies, televised narratives, and edutainment. Distractions can be programmed into these new digital experiences as well. I expect a synthesis of all this theoretical work as soon as the novelty of the new research topics has receded, and we will begin to see systematic exploration of how to suppress counterarguing across a broader set of communication experiences. As new media and technologies appear, we can expect that social scientists will continue the patterns of persuasion research that we have seen. That research prizes effectiveness and respects only the most obvious ethical limits. Suppression of critical thinking is too nuanced an ethical consideration for their research programs.

8.5 Community Argument? Concerns and Ambitions Community arguing is an ideal, and much of the case I have made for it can be found in prior work by van Eemeren and Grootendorst (2004), Habermas (reviewed by Burleson & Kline, 1979), Johnstone (1982), and Wenzel (1980). From one perspective or another, all of these scholars have demonstrated that good arguing requires that the arguers be equally advantaged: they should have the same status and power during the discussion, the same rights, the same ability if that is feasible, and the same resources. The arguments themselves need to be equally available as well. People must understand that someone is trying to persuade them, and be able to examine the persuader’s case. These last two points have always been obvious, but seem to be taking on new force in our new century. It is not an ideal that is easily attainable, at least in the near-perfect sense that one might wishfully write about. Humans have evolved to be unfriendly to outsiders’ influence, and we ignore it or aggressively reject it out of hand. Neither silence nor word-throwing involves much high quality arguing. Should we genuinely try to argue properly, we do not remember well enough to bring all our own resources into the argument, even when we are motivated to do so. Fortunately, we have also evolved to argue about our differences, to learn epistemic vigilance, and to make reasoned choices. The trouble is that these evolutionary inheritances are in a balance, both for our societies and for us as individuals. A noticeable quantity of social scientific research is adding weight to one side of that balance, but not the one that we might want. I hope that I have made a case for concern about the ethical implications of persuasion research. I have reviewed some of the literature on distraction, nudges, narrative, and visual appeals to show that suppression of counterarguing is an avowed research objective. The social scientists interested in persuasion only rarely have intellectual backgrounds and commitments that match those of argumentation scholars. Many of us, for example, have been at least tempted by readings in moral philosophy at some

8 Worries About the Prospects for Community Argument

125

point in our careers. We learned about drawing conclusions through exposure to logic and dialectic, rather than from modern investigations of corporate decision-making. We see persuasion as a morals-laden effort to alter another person’s life, rather than something planned by an advertiser or a political consultant with both eyes on effectiveness. And perhaps we might as well admit that as individuals most of us are disinclined to let anything pass by, unexamined. But the scholars whose work I have worried about seem not to have been educated to valorize critical thinking, at least as we understand it, and when they do reflect on the ability to argue, they seem not to notice that this affordance has ethical implications for personal and public life. This research will not stop, because suppressing counterarguing is an established technique for increasing persuasive effectiveness, profits, public compliance, and votes. So this is going to fall to us, to the very specific academic community of argument scholars. I think that we need to develop theories and curricula to teach critical thinking about non-propositional and nonverbal arguments. We have had some modest success in teaching critical thinking (Abrami et al., 2008), as conceptualized by our own Delphi report on critical thinking (Facione, 1990). Students’ improvements were most marked in elementary and middle school, with quite small effects reported for high school and college (Abrami et al., 2008, p. 1117). Huber and Kuncel (2016) gave a more positive report about university education’s effects on critical thinking, but they reviewed the ordinary experience of college rather than focused pedagogy. However, as we are well aware, these conceptualizations, interventions, and measurements of critical thinking all presuppose linear verbal arguments. We must develop theories and then pedagogies for doing at least as well for nonverbal materials as we do for verbal ones. Presently, I think our best start is to be found in our literatures about visual and narrative argument. Godden (2013) argued that visual arguments can be successfully analyzed by means of the theories we have developed for verbal arguments, and others have pursued that same idea. Ripley (2008), for instance, analyzed a simple advertisement for ceramic tile of the sort one might install in a bathroom, and systematically connected that message to theories from formal logic, informal logic, the Toulmin model, multimodal argumentation, and pragma-dialectics. Dove (2016) applied various argument schemes and their allied critical questions to a simple graphic illustration of the domino theory used by the U.S. to justify military action in Vietnam. The work to date persuades me that we can analyze visual arguments using verbal tools, but I think whether those theories can give a satisfactorily complete analysis of visual arguments remains an open question. For one thing, we will need to be quite clear about when one of these visual events is actually intended to be persuasive, because the world is full of innocent distractions, TV shows, and doodles. In the realm of narrative persuasion, Fisher’s (1987) focus on internal and external consistency seem useful. In fact, these two ideas seem analogous to me to the illative core and dialectical tier that we have met elsewhere (Johnson, 2000). It might be possible to apply Fisher’s ideas to visual arguments of some sorts, and certainly to edutainment.

126

D. Hample

I have some tentative suggestions as to how we might teach these sorts of things to children, then young adults. Classroom teachers should decide what things ought to be done at each age level, but here is a draft of pedagogical stages in teaching how to think about various messages: 1. What is the point of this story, letter, song, game, or cartoon? Children might benefit by seeing examples that don’t actually have a persuasive point, like an entertaining board game or playing dolls. 2. Does someone really want you to accept their point? Why do you think they want that? a. Does that person have the right to change what you think or how you act? Obedience is part of a child’s life, but the force of simple obedience changes with age and family, and the invitation to argue probably does as well. b. Does that person give you reasons to change? The reasons might be stated or they may be an implicit part of the point. Maybe they need to be figured out, just as an enthymeme has to be completed to understand it. Are the reasons good ones? What should we think about a persuasion message that doesn’t seem to have reasons? c. Is there anything in the message that doesn’t seem necessary to the point? If so, why do you suppose it is there? Extra materials might be for pure entertainment, or they might be distractions. 3. Why does the message just seem right to you, or why does it just seem wrong? For older students, this would be the chance to talk about our desires for sociality and defense of in-groups. Should you try to discipline yourself to reverse your natural reaction? When is that justified, and when is it a waste of energy? Frankly I am doubtful that I am the person to design such a years-long curriculum, but perhaps this can serve as a platform for development or discussion. One can see where to fit such basic matters as fact versus opinion (#1), the way the content and source combine to create meaning (#2), fallacies (#2b), evidence (#2b), distraction (#2c), and how we naturally react to agreement and disagreement (#3). Scholarly approaches such as Fisher’s and the others just mentioned can be fit into the system depending on what kind of message form is being studied. Teachers might like to teach linear verbal messages, stories, poems and songs, drawn or animated cartoons, child-targeted edutainments, video games, and the other tactics considered in this essay, in age-appropriate order. If it has a point, we can learn to find it, and then to think about whether to accept it. At a minimum, people need to learn how to identify those features of an argument that are intended to hide it—that is, to suppress counterargument or other intellectual engagement. We must be able to notice dangerous elements in an argument (“Did you hear how the music rises up to distract you during the recitation of the drug’s side effects”), and we ought to figure out how to detect the targeting of certain audiences (“Why does this edutainment drama feature attractive vampires?”), as is standard in propaganda analysis (Jowett & O’Donnell, 2019). These are not the sorts of things

8 Worries About the Prospects for Community Argument

127

that we ordinarily think about in terms of argumentation theory. Perhaps we need to be more practical and absorb more social science findings into our work, in order to promote critical thinking about new communication experiences. I believe that we need to develop something akin to the old Institute for Propaganda Analysis program for K-12 education (Sproule, 1997). They came up with a list of about half a dozen fallacies, gave them bright contemporary names, and taught students to recognize them. That approach, of course, now seems to have been unduly restricted to verbal appeals. Our community needs to develop a simple list of basic ideas that apply across the board—to verbal arguments, but also to visual ones, musical appeals, stories, virtual reality, digital games, product placements, nudges, distractions, and all the rest. It is, as many of us teach in our classes, a matter of public self-defense.

References Abrami, P. C., et al. (2008). Instructional interventions affecting critical thinking skills and dispositions: A stage 1 meta-analysis. Review of Educational Research, 78, 1102–1134. Bateson, P. (1988). The biological evolution of cooperation and trust. In D. Gambetta (Ed.), Trust: Making and breaking cooperative relations (pp. 14–30). Oxford: Blackwell. Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for interpersonal attachments as a fundamental human motivation. Psychological Bulletin, 117, 497–529. Beauvoir, S. D. (1972). The second sex (H. M. Parshley, Trans.). Harmondsworth: Penguin. BETA [Behavioural Economics Team of the Australian Government]. (2019). Behavioral economics. https://www.pmc.gov.au/domestic-policy/behavioural-economics. Retrieved March 20, 2019. Bilandzic, H., & Busselle, R. (2013). Narrative persuasion. In J. P. Dillard & L. Shen (Eds.), The SAGE handbook of persuasion: Developments in theory and practice (2d ed., pp. 200–219). Los Angeles, CA: SAGE. Birdsell, D. S., & Groarke, L. (1996). Toward a theory of visual argumentation. Argumentation and Advocacy, 33, 1–10. Blascovich, J., Loomis, J., Beall, A. C., Swinth, K. R., Hoyt, C. L., & Bailenson, J. N. (2002). Immersive virtual environment technology as a methodological tool for social psychology. Psychological Inquiry, 13, 103–124. Boehm, C. (1999). Hierarchy in the forest: The evolution of egalitarian behavior. Cambridge, MA: Harvard University Press. Brockriede, W. (1975). Where is argument? Journal of the American Forensic Association, 11, 179–182. Buller, D. (1986). Distraction during persuasive communication: A meta-analytic review. Communication Monographs, 53, 91–114. Burleson, B. R., & Kline, S. L. (1979). Habermas’ theory of communication: A critical explication. Quarterly Journal of Speech, 65, 412–428. Campbell, G. (1988). The philosophy of rhetoric. Carbondale, IL: SIU Press. Campbell, D. (2017). Cleverer than command. Social & Legal Studies, 26, 111–126. Canary, D. J., Brossmann, J. E., Brossmann, B. G., & Weger, H., Jr. (1995). Toward a theory of minimally rational argument: Analyses of episode-specific effects of argument structures. Communication Monographs, 62, 183–212. Carson, R. (1962). Silent spring. New York, NY: Houghton Mifflin. Cherniak, C. (1986). Minimal rationality. Cambridge, MA: MIT Press.

128

D. Hample

Cho, J., Ahmed, S., Keum, H., Choi, Y. J., & Lee, J. H. (2018). Influencing myself: Selfreinforcement through online political expression. Communication Research, 45, 83–111. Cialdini, R. (1984). Influence: The psychology of persuasion. New York, NY: Morrow. Cin, S. D., Zanna, M. P., & Fong, G. T. (2004). Narrative persuasion and overcoming resistance. In E. S. Knowles & J. A. Linn (Eds.), Resistance and persuasion (pp. 175–191). Mahwah, NJ: Erlbaum. Cosmides, L., & Tooby, J. (1992). Cognitive adaptations for social exchange. In J. H. Barkow, L. Cosmides, & J. Tooby (Eds.), The adapted mind: Evolutionary psychology and the generation of culture (pp. 163–228). New York, NY: Oxford University Press. Crowther-Heyck, H. (2005). Herbert A. Simon: The bounds of reason in modern America. Baltimore, MD: Johns Hopkins University Press. Dove, I. J. (2016). Visual scheming: Assessing visual arguments. Argumentation and Advocacy, 52, 254–264. Edwards, M. U., Jr. (2004). Printing, Propaganda, and Martin Luther. Minneapolis, MN: Fortress Press. Facione, P. (1990). Critical thinking: A statement of expert consensus for purposes of educational assessment and instruction (The Delphi Report). ERIC document ED 315 423. Fagan, B. (2010). Cro-Magnon: How the ice age gave birth to the first modern humans. New York, NY: Bloomsbury Press. Festinger, L., & Maccoby, N. (1964). On resistance to persuasive communications. The Journal of Abnormal and Social Psychology, 68, 359–366. Fisher, W. R. (1987). Human communication as narration: Toward a philosophy of reason, value, and action. Columbia, SC: University of South Carolina Press. Fodor, J. A. (1983). Modularity of mind: An essay on faculty psychology. Cambridge, MA: MIT Press. Gervais, B. T. (2015). Incivility online: Affective and behavioral reactions to uncivil political posts in a web-based experiment. Journal of Information Technology & Politics, 12, 167–185. Gilbert, M. A. (1997). Coalescent argumentation. Mahwah, NJ: Erlbaum. Gilbert, M. A. (2014). Arguing with people. Tonawanda, NY: Broadview Press. Godden, D. M. (2013). On the norms of visual argument. In D. Mohammed, & M. Lewi´nski (Eds.), Virtues of Argumentation. Proceedings of the 10th International Conference of the Ontario Society for the Study of Argumentation (OSSA), 22–26 May 2013 (pp. 1–13). Windsor, ON: OSSA. Groarke, L. (1996). Logic, art and argument. Informal Logic, 18, 105–129. Grunwald, M. (2009, April 2). How Obama is using the science of change. Time. http://www.time. com/time/magazine/article/0,9171,1889153,00.html. Guadagno, R. E., Blascovich, J., Bailenson, J. N., & McCall, C. (2007). Virtual humans and persuasion: The effects of agency and behavioral realism. Media Psychology, 10, 1–22. Habermas, J. (1989). The structural transformation of the public sphere: An inquiry into a category of bourgeois society (T. Burger, Trans.). Cambridge, MA: MIT Press. Halpern, D. (2015). Inside the nudge unit. London: W. H. Allen. Hample, D. (2009). Argument: Its origin, function, and structure. In D. S. Gouran (Ed.), The functions of argument and social context (pp. 1–10). Alta, UT: National Communication Association. Hample, D. (2018). Interpersonal arguing. New York, NY: Peter Lang. Hample, D., Han, B., & Payne, D. (2010). The aggressiveness of playful arguments. Argumentation, 24, 405–421. Holt-Lunstad, J., Smith, T. B., Baker, M., Harris, T., & Stephenson, D. (2015). Loneliness and social isolation as risk factors for mortality: A meta-analytic review. Perspectives on Psychological Science, 10, 227–237. Huber, C. R., & Kuncel, N. R. (2016). Does college teach critical thinking? A meta-analysis. Review of Educational Research, 86, 431–468. Jackson, S. (2017, June). Naturally occurring argumentation. Paper presented at the meeting of the European Conference on Argumentation, Fribourg, Switzerland. Jensen, J. V. (1983). Metaphor in argumentation. Rhetoric Society Quarterly, 13, 201–207.

8 Worries About the Prospects for Community Argument

129

Jeong, S.-H., & Hwang, Y. (2012). Does multitasking increase or decrease persuasion? Effects of multitasking on comprehension and counterarguing. Journal of Communication, 62, 571–587. Johnson, R. H. (2000). Manifest rationality: A pragmatic theory of argument. Mahwah, NJ: Lawrence Erlbaum. Johnson, E. J., et al. (2012). Beyond nudges: Tools of a choice architecture. Mark Lett, 23, 487–504. Johnstone, H. W., Jr. (1982). Bilaterality in argument and communication. In J. R. Cox & C. A. Willard (Eds.), Advances in argumentation theory and research (pp. 95–102). Carbondale, IL: Southern Illinois University Press. Jowett, G. S., & O’Donnell, V. (2019). Propaganda & Persuasion (7th ed.). Los Angeles, CA: Sage. Kenrick, D. T. (2012). Evolutionary theory and human social behavior. In P. A. M. Van Lange, A. W. Kruglanski, & E. T. Higgins (Eds.), Handbook of theories of social psychology (Vol. 1, pp. 11–31). Los Angeles, CA: Sage. Lake, R. A. (1983). Enacting red power: The consummatory function in Native American protest rhetoric. Quarterly Journal of Speech, 69, 127–142. Ma, Z. (2019). Effects of immersive stories on prosocial attitudes and willingness to help: Testing psychological mechanisms. Media Psychology. Online first publication. https://doi.org/10.1080/ 15213269.2019.16516. Mackie, D. M., Devos, T., & Smith, E. R. (2000). Intergroup emotions: Explaining offensive action tendencies in an intergroup context. Journal of Personality and Social Psychology, 79, 602–616. Matthes, J., Knoll, J., & von Sikorski, C. (2018). The “spiral of silence” revisited: A meta-analysis on the relationship between perceptions of opinion support and political opinion expression. Communication Research, 45, 3–33. Mercier, H., & Sperber, D. (2017). The enigma of reason. Cambridge, MA: Harvard University Press. Molina, R. G., & Jennings, F. J. (2018). The role of civility and metacommunication in Facebook discussions. Communication Studies, 69, 42–66. Moyer-Gusé, E. (2008). Toward a theory of entertainment persuasion: Explaining the persuasive effects of entertainment-education messages. Communication Theory, 18, 407–425. Moyer-Gusé, E., & Nabi, R. L. (2010). Explaining the effects of narrative in an entertainment television program: Overcoming resistance to persuasion. Human Communication Research, 36, 26–52. Niederdeppe, J., Shapiro, M. A., & Porticella, N. (2011). Attributions of responsibility for obesity: Narrative communication reduces reactive counterarguing among liberals. Human Communication Research, 37, 295–323. Noelle-Neumann, E. (1974). The spiral of silence: A theory of public opinion. Journal of Communication, 24, 43–51. Olmos, P. (2013). Narration as argument. In D. Mohammed, & M. Lewi´nski (Eds.), Virtues of Argumentation. Proceedings of the 10th International Conference of the Ontario Society for the Study of Argumentation (OSSA), 22–26 May 2013 (pp. 1–14). Windsor, ON: OSSA. Petty, R. E., Wells, G. L., & Brock, T. C. (1976). Distraction can enhance or reduce yielding to propaganda: Thought disruption versus effort justification. Journal of Personality and Social Psychology, 34, 874–884. Quine, W. V., & Ullian, J. S. (1978). The web of belief (2nd ed.). New York, NY: Random House. Ripley, M. L. (2008). Argumentation theorists argue that an ad is an argument. Argumentation, 22, 507–519. Roos, P., Gelfand, M., Dau, D., & Lun, J. (2015). Societal threat and cultural variation in the strength of social norms: An evolutionary basis. Organizational Behavior and Human Decision Process, 127, 14–23. Santabañez, C., Hample, D., & Hample, J. M. (2019, June). How do Chilean seniors think about arguing? Paper presented at the meeting of the European Conference on Argumentation, Gröningen, The Netherlands.

130

D. Hample

Shen, L., & Bigsby, E. (2013). The effects of message features: Content, structure, and style. In J. P. Dillard & L. Shen (Eds.), The SAGE handbook of persuasion: Developments in theory and practice (2d ed., pp. 20–35). Los Angeles, CA: SAGE. Shen, F., & Han, J. (2014). Effectiveness of entertainment education in communicating health information: A systematic review. Asian Journal of Communication, 24, 605–616. Shen, L., & Seung, S.-Y. (2018). On measures of message elaboration in narrative communication. Communication Quarterly, 66, 79–95. Simons, J. J. P., & Green, M. C. (2018). Divisive topics as social threats. Communication Research, 45, 165–187. Simpson, J. A., & Belsky, J. (2016). Attachment theory within a modern evolutionary framework. In J. Cassidy & P. R. Shaver (Eds.), Handbook of attachment: Theory, research, and clinical applications (pp. 91–116). New York, NY: Guilford. Slater, M. D. (1997). Persuasion processes across receiver goals and message genres. Communication Theory, 7, 125–148. Slater, M. D., & Rouner, D. (2002). Entertainment-education and elaboration likelihood: Understanding the processing of narrative persuasion. Communication Theory, 12, 173–191. Snow, C. P. (1963). Two cultures. New American Library. Social and Behavioral Sciences Team. (2015). Annual report. https://sbst.gov/download/2015% 20SBST%20Annual%20Report.pdf. Spoor, J. R., & Kelly, J. R. (2004). The evolutionary significance of affect in groups: Communication and group bonding. Group Processes and Intergroup Relations, 7, 398–412. Sprain, L., & Black, L. (2018). Deliberative moments: Understanding deliberation as an interactional accomplishment. Western Journal of Communication, 82, 336–355. Sproule, J. M. (1997). Propaganda and democracy: The American experience of media and mass persuasion. Cambridge, MA: Cambridge University Press. Stroeker, N. E. (2016). An overview of behavioral economics in Dutch policy making. The next step: How to nudge policy makers. Applied Studies in Agribusiness and Commerce—ABSTRACT, 10, 27–32. Taylor, S. E. (2012). Tend and befriend theory. In P. A. M. Van Lange, A. W. Kruglanski, & E. T. Higgins (Eds.), Handbook of theories of social psychology (Vol. 1, pp. 32–49). Los Angeles, CA: Sage. Thaler, R. H., & Sunstein, C. R. (1999). Nudge: Improving decisions about health, wealth, and happiness. New Haven, CT: Yale University Press. van Eemeren, F. H., & Grootendorst, R. (2004). A systematic theory of argumentation: The pragmadialectical approach. Cambridge, MA: Cambridge University Press. van Eemeren, F. H., Grootendorst, R., Jackson, S., & Jacobs, S. (1993). Reconstructing argumentative discourse. Tuscaloosa, AL: University of Alabama Press. Waiguny, M. K. J., Nelson, M. R., & Terlutter, R. (2012). Entertainment matters! The relationship between challenge and persuasiveness of an advergame for children. Journal of Marketing Communications, 18, 69–89. Watzlawick, P., Beavin, J. H., & Jackson, D. D. (1967). Pragmatics of human communication. New York, NY: Norton. Weger, H., Jr., & Canary, D. J. (2010). Conversational argument in close relationships: A case for studying argument sequences. Communication Methods and Measures, 4, 65–87. Wenzel, J. W. (1980). Perspectives on argument. In J. Rhodes, & S. Newell (Eds.), Proceedings of the Summer Conference on Argumentation (pp. 112–133). Alta, UT: Speech Communication Association. Wrangham, R. (2019). The goodness paradox: How evolution made us more and less violent. London: Profile Books.

Chapter 9

Assessing Connection Adequacy for Arguments with Institutional Warrants James B. Freeman

9.1 Warrants, Sources of Backing, and Rules An institutional warrant is backed by the rules of some institution, such as a branch of law or a particular game. Such a warrant is somehow “derivable” from the rules of the backing institution, as one might argue “down” from premises to conclusion in a derivation. By contrast, a warrant is “backed from below” when the backing consists ultimately of perceptually observed evidence. The backing is built up inductively from this body of evidence. The law is a paradigm case of an institution, but as Searle points out in (1969, p. 51), there are many other institutions, games being prime illustrations. Also, certain human relations such as marriage presuppose institutions to be entered into. In this paper, we shall develop our account with respect to warrants backed by laws, and then point out how our conclusions may be applied to institutional warrants in general. We may distinguish two types of rules, both included in bodies of law. Cohen in (1970, p. 157) considers legal hypotheses having the following form: For any persons x and y, if x has R to y, then x has good cause of action against y (i.e. if x sues y x ought to win).

(1970, p. 157) This hypothesis and the law from which it is derived is regulative. It specifies that certain conditions where two persons are related in a particular way have legal consequences. The rule assumes that we can recognize when this relation holds. We do not need this rule for the condition to be defined or to exist. By contrast, constitutive rules are necessary for certain conditions to obtain or for certain properties to be satisfied. A constitutive rule has the form “X counts as Y in context C” (Searle, 1969, p. 52). For example, a piece of paper of a certain composition and bearing certain marks counts as a $10 bill, given the declaration of J. B. Freeman (B) Department of Philosophy, Hunter College of the City University of New York, New York, USA e-mail: [email protected] © Springer Nature Switzerland AG 2020 F. H. van Eemeren and B. Garssen (eds.), From Argument Schemes to Argumentative Relations in the Wild, Argumentation Library 35, https://doi.org/10.1007/978-3-030-28367-4_9

131

132

J. B. Freeman

such a type of object, of such a composition, bearing such a pattern of marks being a $10 bill by a duly authorized government body. As Searle would emphasize, one cannot completely describe the object as a $10 bill without the constitutive currency rules. How do rules of either type back warrants? We regard warrants as licences to move from a set of premises to a conclusion. Consider Toulmin’s paradigm example: From:

x was born in Bermuda

To infer: x is a British subject

Why, Toulmin invites us to ask, does the warrant have “authority and currency”? (1958, p. 103) The warrant is supported by backing evidence. In the empirical case, the body of evidence consists of observed instances of co-variation. If elements over which the variables of the rule range satisfy the property or relation indicated by the premise of the rule, they satisfy the property or relation indicated by the conclusion. In (2005, pp. 136–37), we pointed out that there are two ways in which one may come to know the rules upon which institutional warrants are formed: either informally through practice or formally through encountering some official or authoritative formulation of the rule. Informal learning is the primary way one may come to learn rules. Coming to recognize some rule by consulting some certified source presupposes that one understands the language used to communicate the rule. But language is a system of syntactic, semantic, and pragmatic constitutive rules. One learns one’s first language informally, by “picking up” its rules, in particular vocabulary and grammatical rules. What we have said about institutional rules applies straightforwardly to institutional warrants. Learning constitutive rules involves learning to make certain projections (i.e. inferences). Upon observing that a soccer ball has taken a certain trajectory, one projects that one has scored a goal. If one has learned a system of rules constituting some practice through engaging in that practice or observing others engaging in that practice, the test of one’s having learned the rule is projecting competently from one’s practice or observation. If the fans erupt cheering, one needs no further confirmation of one’s projection that a player has scored a goal or of one’s having learned correctly the rule behind it, at least for this application. A correctly learned rule backs a reliable warrant. Likewise, learning regulative rules involves learning to make projections. From the fact that Jones has damaged Smith’s property, we learn to project that Jones owes damages to Smith.

9.2 From Bodies of Law to Formally Backed Warrants Getting from laws which back warrants to the warrants themselves involves more steps than getting from the backing of empirical warrants to the warrants backed. In the empirical case, one goes from a body of observations or observation reports–our own or others–to the warrant. In the simplest case, unaided perceptual observations constitute the backing. Other cases may involve what we called in (2005, pp. 296–97) institutional testimony. Simple perception has been interpreted in accord with some

9 Assessing Connection Adequacy for Arguments …

133

institution. For example, what does a trained x-ray technician “see” when viewing an x-ray? Besides patterns of grey, the technician may see something medically significant, such as a broken bone or a tumor. The point is that the technician must be trained to receive this information by interpreting the brute (Compare Searle, 1969, p. 50) information from the x-ray according to the institution of interpreting x-rays. Likewise, going from a written transcript of one or more enacted laws to the backing it may provide for a warrant involves institutional interpretation, in particular interpretation according to rules of the language of the transcript. One must somehow perceive a transcript to be able to understand it. But one must also understand how to get from the written transcript of a law (or body of laws) to the warrant and how one may do that reliably. We propose a four-step procedure. Recall that a warrant is a licence to proceed from one or more premise schemes to a conclusion scheme. We say schemes because a warrant will contain individual variables or schematic letters, indicating its generality. To simplify our discussion, let us set schematic letters aside. In the first step of the procedure, one must go from the transcript of the law to the transcript as understood. Second, one must express that understanding as an A or E categorical. Third, where the subject of the categorical may be paraphrased according to the form ‘x1 , . . . , xn ’ and the predicate as ‘x1 , . . . , xn ’, one transforms the statement into (∀x1 ) . . . (∀xn )(x1 , . . . , xn ⊃ x1 , . . . , xn ) For example, All citizens of the United States are entitled to vote in national elections

becomes (∀x)(x is a citizen of the United States ⊃ x is entitled to vote in national elections)

Likewise Should one person damage another’s person’s property, he owes that person reparations

becomes (∀x)(∀y)([x damages y’s property & x = y] ⊃ x owes reparations to y)

Fourth and finally, one transforms the (partially symbolized) categorical statements into a warrant by dropping the initial quantifiers, making the antecedent of the conditional the premise (schema) of the warrant and the consequent the conclusion schema. Thus our quasi-symbolic paraphrases become From:

x is a citizen of the United States

To infer: x is entitled to vote in national elections

and From

x damages y’s property & x = y

134

J. B. Freeman

To infer: x owes reparations to y

The first step in this procedure requires some explication. What is involved in understanding a transcript of a law? We address that question in the next section.

9.3 Reliably Understood Transcripts and Speech Acts A warrant ultimately backed by a transcript of a law or other legal principle is reliable when it expresses a correctly drawn inferential moral from that transcript. When is there a presumption that we have drawn the moral correctly? We approach this question through Searle’s speech act theory. Suppose the formulation we consult is included in the official published record of the acts of some legislative body. We are dealing here with two speech acts in tandem. The formulation is an assertive that certain rules containing certain provisions have been enacted. These assertives report that another speech act has taken place, a declaration by a formal legislative body. Searle defines a declaration as an illocutionary act whose successful performance “brings about the correspondence between the propositional content and reality, successful performance guarantees that the propositional content corresponds to the world” (1979, pp. 16–17). Thus, if the official rule keepers of baseball have declared that balls hit with a certain trajectory are foul, balls hit with that trajectory are foul. For the derived warrant to be presumptively reliable, four conditions are necessary: (1) One must understand the assertion that a successful declaration has been made. (2) That assertion must be true. (3) One must understand the propositional content of the declaration. (4) That content must give the warrant authority or currency. (Notice that if one heard the declaration of the rule from the declaring authority directly, as opposed to encountering some assertion of that declaration, to form a warrant properly one would still have to understand both that the authoritative body had made a declaration and what was the propositional content of that declaration. Again that content must authorize the warrant and thus, as we discussed in the last section, it must be in the form of a universally generalized conditional.) Under what circumstances will these four conditions be satisfied? That an assertion is true is simply a matter of how the world is. The second condition, then, is totally straightforward. What is involved in the issue of understanding that an assertive has been made to the effect that a declaration with a certain propositional content has been made? Here again we may turn to Searle’s speech act analysis. We may factor a sentence expressing a declarative into two parts, a declarative illocutionary force indicator D, such as “I declare” or “It is being declared that” and a sentence P expressing a proposition that p, the proposition declared. One understands a sentence of the form DP then when one knows the rules for the declarative operator and the rules for understanding P. In (2005), patterned on Searle’s account of rules for the illocutionary promise operator, we presented four rules for using D: Rule 1. D is to be uttered only in the context of a sentence P used to express p. Rule 2. D is to be uttered only if that p does not occur in the normal course of events.

9 Assessing Connection Adequacy for Arguments …

135

Rule 3. D is to be uttered only by a speaker in a position of legislative authority. Rule 4. Uttering D in the context of P brings it about that p (2005, p. 198; compare Searle, 1969, p. 63.) According to Searle, one understands a sentence P when one knows its meaning, which is determined by rules specifying both conditions of utterance of the sentence and what the utterance counts as (1969, p. 48). Clearly, a sentence with the illocutionary force of an assertive that a declarative has been made is of the form ADP, where A is the assertive operator “I assert” or “It is asserted that.” Understanding such a sentence, in addition to understanding DP is a matter of knowing the rules for using A. Hence, properly acquiring a warrant when confronted either with the declaration of a rule or an assertive to the effect that such a rule has been declared requires understanding both the propositional content of the rule, the rules for the proper use of the declarative operator and, in the case of assertives, for the assertive operator as well. What then is involved in understanding propositional content? Searle presents his account of sentence meaning in (1969, especially pp. 48–50). According to him, “Understanding the speaker’s utterance is closely connected with recognizing his intentions,” more specifically “to get the hearer to know (recognize, be aware of) that certain states of affairs specified by certain of the rules [of the language common to speaker and hearer] obtain” (1969, p. 48). According to Searle, understanding the meaning of P requires that the speaker mean that P. This involves three components: (a) intending (i-i) to get the hearer to know (recognize, be aware of) that certain states of affairs specified by certain of the rules obtain [in the case of declaratives to get the hearer to recognize that these states of affairs obtain by virtue of the sentence being declared], (b) intending to get the hearer to know (recognize, be aware of) these things by getting him to recognize (i-i) and (c) intending to get him to recognize (i-i) in virtue of his knowledge of the rules for the sentence uttered …. The hearer’s understanding the utterance will simply consist in those intentions being achieved. (1969, p. 48)

Searle presents rules for the meaning of referring expressions and predicating expressions which may be constituents of sentences in (1969, pp. 94–96 and 126–27). Institutional declaratives, in particular legal declaratives, will also involve quantifier expressions and sentential connectives. Presenting Searle’s rules for reference and predication and supplementing them with rules for quantifier and connective meaning is beyond the scope of this paper. Searle’s account of sentence meaning is sufficient for understanding how we may get from the transcript of a rule or law to the rule or law as understood. We must point out here that Searle intends this procedure to result in understanding the literal meaning of what has been uttered. A fuller understanding of the meaning may involve a further “reading” of the intentions of those who uttered the declarative–what they have meant or not meant by this declaration. We shall discuss this problem shortly. Step 2 of the procedure of going from the institutional rule to the warrant it backs may in a number of cases be totally straightforward. Laws may be expressed as A or E categoricals. For example

136

J. B. Freeman

(1) Every person … or other legal entity contracting with any person … accused or convicted of a crime in this state, with respect to the reenactment of such crime [through the media] … shall submit a copy of such contract to the board and pay over to the board any moneys which would otherwise, by terms of such contract, be owing to the person so accused or convicted or his representatives. N. Y. Exec. Law § 632 This law was enacted by the New York State Legislature to prevent David Berkowitz (alias Son of Sam), a serial murderer, to benefit from the sale of his story. In a series of crimes, Berkowitz left five dead and others injured. Although complex, this statement is clearly an A-categorical. (2) No person in the United States shall, on the ground of race, color, or national origin, be excluded from participation in, be denied the benefits of, or be subject to discrimination under any program or activity receiving Federal financial assistance. Title VI of the Civil Rights Act of 1964 (3) To recover for negligence, the plaintiff must establish each of the following elements: duty, standard of care, breach of duty, cause-in-fact, proximate cause (scope of liability) and damages. Passgraf v. Long Island Railroad Company Although this statement is not in the explicit form of a categorical, paraphrasing it in the form of an A-categorical is totally straightforward: All plaintiffs seeking to recover for damages are persons required to establish duty, standard of care, breach of duty, cause-in-fact, proximate cause, and damages.

Although paraphrasing may be necessary, this statement is clearly an A-categorical. Step 3 calls for a quasi-symbolism of the paraphrase. In approaching this step, some may actually want to symbolize the statement in a first-order predicate language or some extension of such a language. We do not believe introducing symbolism for term or predicate expressions is absolutely necessary, although it may be convenient. However, the step prior to actually symbolizing, paraphrasing the sentence in one’s mind to identify the specific components of the sentence and how they fit together, may be practically necessary for a proper symbolism and is what Step 3 calls for. Let’s apply this step to quasi-symbolizing the second law we just noted: For all x, if x is a person in the United States, then it is not the case that there is a y, y is a program and y receives Federal financial assistance and x is excluded from participating in y on the grounds of race, color, or national origin or x is denied the benefits of y on the ground of race, color, or national origin, or x is subject to discrimination in y on the ground of race, color, or national origin.

The structure of this sentence as a universally quantified conditional is now plain. Step 4 is now trivial. There is one initial universal quantifier. Drop it, take the antecedent as the premise of the warrant and the consequent as the conclusion. We have now shown how one can get from an inscribed rule to the warrant it backs. The question now is when does the backing of the rule render the warrant a reliable licence to move from an instance of the premise to an instance of the conclusion. We turn to that question in the next section.

9 Assessing Connection Adequacy for Arguments …

137

9.4 Reliable Institutional Warrants, Rebuttals, and Argument Cogency For an institutional warrant to be reliable, it must be true that a body has declared the backing rule, that body has the legitimate authority to declare the rule, and the propositional content of the rule is properly understood. Recall that for the premises of an argument to be adequately connected to the conclusion, they must be relevant to the conclusion and they must constitute grounds adequate for accepting the conclusion should the premises be acceptable. Should an institutional warrant be reliable, then where it licences a move from premises to conclusion, the premises are ipso facto relevant to the conclusion. As we discussed in (2013, pp. 31–33), we may divide arguments into conclusive and defeasible. We may likewise divide warrants into conclusive and defeasible. If the institutional rule backing the warrant is without exception, then the warrant is conclusive. For example, consider this formulation for the rule of client privilege: No lawyer may be required or compelled to disclose any statement a client has made to him or her.

If this law is literally absolute, then the warrant From:

x is a lawyer and y is a client of x and y has told z to x

To infer: it is not the case that x can be required to disclose z

Any instance of this rule is a conclusive argument. But laws may not be intended to be absolute. Certain exceptions may be explicitly or implicitly allowed. Further, terms in law may be vague, legislators leaving it to the discretion of the courts to reduce the vagueness. Step 2 of the procedure may not be sufficient to determine the literal meaning of the rule. In addition, the rule as literally interpreted may not be the intended backing of the warrant. At least, it may be argued that the rule should be interpreted non-literally. We shall discuss this case shortly. Toulmin’s example From:

x was born in Bermuda

To infer: x is a British subject

illustrates the first case. Although not explicitly mentioned, there may be ways a person born in Bermuda may still not be a British subject. Neither of the parents were British subjects while British law requires that at least one of the parents be a British subject. Consider the rule All motor vehicles must be licenced.

What is a motor vehicle? The term admits of a range of indefiniteness where whether or not the term applies is questionable. Automobiles, motor cycles, trucks clearly count as motor vehicles. A human powered bicycle does not. But what about a motorized bicycle? Should it require a licence and registration?

138

J. B. Freeman

The case of Riggs v. Palmer illustrates how the proper understanding of a rule may need to go beyond the literal meaning of the law. Francis B. Palmer, in a properly executed will, left his property to his two daughters, Mrs. Riggs and Mrs. Preston, and his grandson, Elmer Palmer, with the largest share going to the grandson. Two years later, to insure the grandfather would not alter his will, Elmer murdered him by poisoning. Should Elmer inherit according to the provisions of the will? The pertinent statute reads this way: No will in writing, except in cases hereinafter mentioned, nor any part thereof, shall be revoked or altered otherwise.

The literal meaning of the statute directs that the grandson inherit the share as stated in the will. But should someone inherit from someone he murdered? Mrs. Riggs sued to prevent that outcome, The court decided for her and against Elmer, reasoning that the legislators could not have intended to allow persons to be able to profit from their crimes. In interpreting the law, they went beyond the literal meaning of the law to recognize the moral sensibilities of its authors. In cases where there are exceptions to the backing rule, the warrant is defeasible and should be modified by a “ceteris paribus” clause or by a list of the recognized defeating conditions. From:

x was left y in z’s will

We may infer ceteris paribus : x will receive y upon probate of z’s will or From :

x was left y in z’s will

We may infer that

x will receive y upon the probate of z’s will unless C1 xyz or C2 xyz or … or Cn xyz

Here C1 , . . . , Cn are the excepting conditions alluded to in the statement of the law. A ceteris paribus clause admits the possibility that all things might not be equal, i.e. that there are certain unspecified rebutting conditions. On this conception, warrant strength, at least the distinction between unequivocal and qualified warrant strength, is a function of the possibility of rebuttals in Toulmin’s sense, as we shall develop in Sect. 9.7. May such a warrant, then, ever reliably convey us from the premises of an argument to its conclusion? As we shall develop in Sect. 9.7, the reliability of a defeasible warrant as an inference rule is a function of the plausibility of its rebuttals. But how, unless they are explicitly mentioned, may we recognize these rebuttals? That is the subject of the next section.

9.5 Recognizing Rebuttals How does one come to identify or recognize the rebuttals associated with a particular institutional warrant? If the warrant is backed by a constitutive rule, this will in some cases simply be a matter of understanding the constitutive rule. The “C” in Searle’s paradigm scheme “X counts as Y in context C” may include a list of rebuttals, as Toulmin’s example concerning birth in Bermuda and British nationality suggests. Our

9 Assessing Connection Adequacy for Arguments …

139

recognizing the rebuttals, then, is a matter of our having properly learned the rule, i.e. of our having sufficiently learned the language in which the rule was formulated, and our intentionally applying what we have learned. Recognizing rebuttals, however, may not always be a matter of simply acquiring or understanding rules. Consider this rule: Two persons who have mutually declared their intent to marry each other before a peace officer and have exchanged marriage vows before that peace officer are legally married.

Now suppose there are cases where two persons have satisfied the antecedent and some further relation between them to some degree or one or both of them satisfy some property to some degree. Suppose that in some of these cases, the courts have ruled the marriage valid and in others invalid. Consider child marriage. Suppose one of the parties to the marriage was only ten years old. Upon consideration, a court of appropriate jurisdiction has ruled the marriage invalid when challenged on the basis of age. As with Riggs v. Palmer, the court ruled that the intent of the law makers was not to allow validity for marriages where one of the parties was clearly incapable of giving informed consent or would be open to exploitation in the marriage. By contrast, another marriage has been challenged on grounds of age where both of the parties were at least sixteen. A court of appropriate jurisdiction has upheld the marriage. The contrasting rulings are sufficient to show that age is a relevant variable for determining the validity of the marriage. There may be other cases where the validity of the marriage has been challenged on grounds other than child marriage. The antecedent of the rule has been satisfied in each case, together with some further property or relation to some degree. Suppose that whether the court rules the marriage valid or invalid depends on the degree to which the property holds. In limiting cases, holding to any degree is sufficient to invalidate the marriage. These properties then are relevant variables for determining whether or not a marriage is valid. To get a comprehensive list of relevant variables, we may need to look beyond just marriage. Marriage is a type of contract. Under what circumstances is a contract invalid? A survey of cases when a contract has been declared invalid will be wider than a survey of marriages which have been declared invalid. It seems quite possible that the survey will reveal relevant variables which might invalidate a marriage beyond what a survey of grounds on which marriages have been invalidated reveals. A judgment of the court is rendered and sets a legal precedent. Identifying the relevant variables then is (or at least may be) a matter of researching records of judgments where the court has considered whether some marriage were invalid and on what grounds and to what extent those grounds held. Interpreting the literal meaning of the law and formulating the warrants it backs proceeds from above. The wider interpretation of the law (and the warrants it backs) proceeds from below by examining legal precedents for sustaining or overturning the law under certain circumstances. A question arises for this account. We are identifying relevant variables on the basis of court decisions, where one degree of a property or relation has been ruled sufficient to invalidate some legal rule, while a different or lesser degree of that property or relation has not. Suppose two courts have given contradictory verdicts.

140

J. B. Freeman

Suppose one court has invalidated the rule on the grounds of several degrees of the property or relation holding, while the other has sustained the rule no matter what degree holds. Is the property or relation a relevant variable or not? We must distinguish two cases. Suppose both courts were of the same level. Surely the courts have not rendered their verdicts at the exact moment. So one court is ignoring the precedent set by the other. It seems that the latter verdict should be set aside. If the first court has invalidated the rule, then this property or relation is a relevant variable. Suppose, by contrast, that the two courts were not of the same level, one higher than the other. Cohen in (1970, p. 158) has indicated a way to deal with this situation. Higher courts may reverse the decisions of lower courts. The evidence concerning precedents sought would be the last ruling of the highest court of appeals to hear the case. If a higher court has ruled that a degree of a certain property or relation invalidates a rule, that property or relation counts as a relevant variable.

9.6 Precedents, Conclusive Versus Defeasible Warrants, and Refining Warrants We have now seen how to proceed from institutional rules, in particular legal rules understood literally, to warrants. We have also seen how legal judgments inferred through warrants can be overturned. The condition cited to justify overturning the judgment consists of instances to some degree of a relevant variable and the relevant variable is a rebuttal. Where a court has overturned a judgment, the decision establishes a precedent. For a given branch of law, the grounds on which a particular law of that type has been overturned can be identified by surveying judicial decisions concerning that branch of law. What is the effect of identifying these conditions on the strength of the support the backing gives to the warrant it supports? Let us remember that where courts have overturned some instances of a relevant variable, they need not have overturned all. Let V1 , . . . , Vn be the relevant variables for a given type of case, say marriage. We may identify m-types of the variants of the variables, each element vi being a variant of Vi . The set of m-tuples of these variables then will specify the possible invalidating conditions for marriage. Now it is possible that when a marriage satisfies certain of these conditions (certain n-tuples of these variables), the marriage is not overturned. Let us also understand then where vi , is a variant of a variable Vi , 1 ≤ i ≤ n, −vi , i.e. the condition that vi does not hold, is also a variant of Vi. If necessary, we may indicate that vi holds by writing + vi . Now let be an n-tuple of variants of . Continuing to assume that V1 , …, Vn are variables relevant to the validity of a marriage, let us assume that when contested, a marriage satisfying some combination of is not overturned when challenged in court. The judgment establishes a precedent for this particular n-tuple. Now consider the warrant From:

x and y have declared intent and exchanged vows before z

9 Assessing Connection Adequacy for Arguments …

141

To infer: x and y are married in the eyes of z

Now suppose, per impossible, our survey of the variables and combinations of variants relevant to the validity of marriage reveals no case where an instance of the premise holds and the conclusion does not, no instance of any circumstance which overturns the marriage. Then the backing for this warrant consists of the laws of marriage and the record of precedents fully backing the warrant. In this case, we may recognize the warrant as conclusive. By contrast, and to be more realistic, let us suppose that under some circumstances, a court has invalidated a marriage. What does this mean for the strength of support the backing gives to the warrant? Clearly, it is no longer conclusive. Clearly also the court will have grounds for its decision, specifically that a particular value of a relevant variable holds or that some combination of values of relevant variables holds. The court decision establishes a precedent for these relevant variables. Their holding, or some combination of variants’ holding, at least if each holds to a certain degree, rebuts the warrant and the argument is not cogent. However, in some circumstances where additional information is available, the warrant can be refined to license a cogent argument. Refining involves conceding that under certain conditions, the warrant is not cogent. The conclusion follows ceteris paribus from the premise unless these conditions hold. Conceding them produces a weaker warrant, since the claim is no longer that the conclusion holds on the basis of the premises alone but only if these conditions do not hold. But we may have information that these rebutting conditions do not hold, information we have called in (1991, pp. 161–65) and (2011, p. 25) counterrebuttals. Consider John Doe and Jane Roe declared intent and exchanged marriage vows before Justice Black. Therefore ceteris paribus unless either John Doe or Jane Roe is not yet 16 years old, they are now married. But both John and Jane are 21 years old.

Following our method in (2011, pp. 12–29), we may diagram the argument this way (Fig. 9.1). Adding the counterrebuttal strengthens the argument. (It is an additional premise.) The argument is still defeasible. Realizing further variants of relevant variables or combinations of such variants, we may still rebut the argument. Of course, if our information would allow us to counter all possible rebuttals, the resulting argument would be conclusive. However, we may not have this information. What may we say about the strength of the argument? Suppose that although some rebutting conditions are identified, we have no court precedent taking them into account. If one or more of them hold, the argument is defeated. If countered, the argument is strengthened. But we do not know which. What may we say of the argument instancing the unrefined defeasible warrant? That is our subject in the next section.

142

J. B. Freeman

Fig. 9.1 Argument with a rebuttal and counterrebuttal

9.7 Defeasible Warrants and Warrant Strength Let V1 , . . . , Vn be the relevant variables which, if a variant of that variable holds, alone or in combination with variants of the other relevant variables, rebut the warrant that from x and y’s engaging in a marriage ceremony, we may take it ceteris paribus that they are married. We have a record of some court decisions declaring a marriage invalid on some of these grounds or if those grounds hold to a certain extent. However, for one or more of these grounds, we have no information one way or the other on whether the ground holds in this case. How then, if at all, may we render a judgment on how strongly the premise supports the conclusion? For concreteness, let us suppose five relevant variables constitute grounds on which a marriage may be declared invalid. Although we do not know in the current case whether any of V1 , . . . , V5 hold to some degree, for each Vi , 1 ≤ i ≤ 5, our record of precedents does indicate in how many cases variants of a given relevant variable occur. Hence, our record of precedents shows which relevant variable has the most occurrences of invalidating variants. We are not distinguishing here cases where the court has cited just one variant of a relevant variable as sufficient to invalidate a marriage and where it has cited a conjunction of variants of variables. Likewise, the record of precedents shows which relevant variable has the second highest number of its variants which invalidate a marriage, and so on down to the variable having the least number of invalidating occurrences. This information then lets us order the relevant variables from most to least of the number of invalidating variants of that variable occurring in our record of precedents. What is the point of this ordering? Cohen speaks directly to this question:

9 Assessing Connection Adequacy for Arguments …

143

We must postulate … that somehow a well-ordering can always be determined for the set of variables that are relevant in a particular branch of law. The main way to do this is obviously to postulate that variables always have the same order in our support- theory as a competent judge would assign them an order of relevance, and that a competent judge looks to past cases to determine that order. If one variable has forced more rules to be qualified than another he will take it to be more relevant. … [I]t is just this criterion of relevance that is most serviceable in establishing how much support exists for a hypothesis. It is obviously more important to discover precedents in each variant of [Vi ] than in each variant of [Vj ], if the former variable has forced more rules to be qualified than the latter. (1970, pp. 158–59)

Given this ordering of the relevant variables, we can now address our question of comparative argument strength. Recall our case John Doe and Jane Roe have declared intent to marry each other and have exchanged marriage vows before Justice Black. Therefore John Doe and Jane Roe are married in the eyes of Justice Black.

The warrant of this argument has the form: From : x1 x2 x3 To infer ceteris paribus ψx1 x2 x3 ,  are ternary relations because premise and conclusion of our example predicate ternary relations. Suppose now that we have two instances of this warrant: a b c Therfore ψa b c and a b c Therefore ψa b c How do we determine which is stronger? If we have no information about values of variables relevant to the issue of the validity of marriage holding in these cases, we cannot make a judgment of argument strength. Suppose by contrast that we have information that in the case of a b c , no degree of V1 , i.e. no variant of this relevant variable holds, while a variant of V1 holds for a b c . Clearly then the first argument is stronger. The second argument is rebutted by V1 holding to some degree. Suppose also that although no variant of V1 holds for the first argument, a value of V2 does, while the situation is reverse for the second argument. The first is still stronger than the second. It is more plausible that an instance of V1 holds than an instance of V2 . Evidence now shows that that no instance of V1 holds for the first argument while an instance of V1 holds for the second. The first argument is not rebutted by the most plausible relevant variable while the second is. Now suppose that a value of V1 rebuts neither the first nor the second argument. But suppose that a value of V2 rebuts the first argument but not the second. The second then is stronger. In this way, then, we can rank the strength

144

J. B. Freeman

of an argument with a legal warrant by the order of the last relevant variable which it survives. Those not surviving the first relevant variable have 0 degree strength. Those surviving the first, but only the first have first degree of strength. Those surviving both the first and second relevant variables have at least second degree strength, and so on down to the fifth variable. How then may we go from this ranking to a judgment about the cogency of an argument? That is our topic in the next section.

9.8 Warrant Strength and Argument Cogency Our ranking of relevant variables also gives us a ranking of the plausibility that a degree of the relevant variable’s holding will either in itself be a rebutting value of the warrant or be a component in a rebutting combination of values. How do we understand plausibility? We have discussed this issue for arguments with empirical warrants in (2018, pp. 79–100). Of the five plausibility factors–relevance, testability, compatibility with established results or previous theories, predictive power, simplicity–we found that only compatibility with previous results had a bearing on the plausibility of whether a value of a relevant variable, itself or in combination with others, will be a rebutting value. Our argument there covers our case here also. Hence, if the record of court cases shows that variants of V1 enter into more invalidating combinations of variants V1 , . . . , V5 than do variants of any other variable, on plausibility considerations we may judge that observation of further combinations of relevant variables would again show that more variants of V1 enter into rebutting combinations of variants than variants of V1 , . . . , V5 . How may we proceed then to make a judgment of the cogency of arguing from abc to abc? Since xyz, xyz are ternary relations, a rebuttal to a warrant licencing inferring from xyz to xyz should predicate a ternary relation. To each Vi , 1 ≤ i ≤ 5, we may associate a set of ternary relations, vi xyz indicating values of Vi . Now suppose a search of court decisions has found 50 cases of attempts to establish an instance of xyz on the basis of the corresponding instance of xyz. We have no information about other cases and the class of cases is open. Suppose that in 26 of these, a court has rejected the argument. Since in better than half the known cases a court has rejected the argument we are not justified in arguing according to the warrant W : From : xyz To infer ceteris paribus : ψxyz Suppose that in 20 of these 26 cases, some degree of V1 xyz held of the instance. By hypothesis about V1 , . . . , V5 , the other relevant variables have occurrences in fewer instances in these rebutting cases. Let’s assume that a degree of V2 xyz holds in 15 of these cases, V3 in 10, V4 in six, and V5 in three. Projecting then from examined cases to unexamined, we infer that a variant of V1 holds in 40% of the cases, V2 in 30%,

9 Assessing Connection Adequacy for Arguments …

145

V3 in 20%, V4 in 12%, and V5 in 6%. When are we justified in arguing according to the warrant? I believe we may argue this way: Of the known cases where a court has rejected the argument, by doing a little arithmetic we find that in approximately 77% the overturning has occurred when a degree of V1 holds, approximately 58% for V2 , approximately 39% for V3 , approximately 23% for V4 , and approximately 12% for V5 . It would seem then that in applying W to an unrecorded case, a counterrebuttal to the claim that a variant of V1 or V2 holds is required to argue cogently from an instance of xyz to xyz ceteris paribus. Clearly without such a counterrebuttal, the conclusion has not been established beyond a reasonable doubt. Where no counterrebuttal concerning V3 , V4 , or V5 has been offered, whether an argument instancing W is cogent may depend on the context. How strong a case is required? If V1 , V2 have been countered, then the argument will fail in at most 39% of the cases. We might claim that the conclusion holds on preponderance of evidence, even when a variant of V3 might hold. We may argue similarly for V4 and V5 . The case of V5 at least, and perhaps for V4 also, is interesting. Should we know that no variant of V1 , V2 , V3 holds, but we do not know this for V5 , could we argue according to the warrant? Since our information on judicial conditions indicates that in only approximately 12% of the cases where an argument is rejected is a variant of V5 involved, how seriously affected is W by this lack of knowledge of whether some value of V5 holds? It is implausible that it does. In the case of V5 , we may say that the warrant is rebuttal tolerant. Even without a counterrebuttal, the argument is cogent. Similar considerations may apply to V4 .

9.9 Application to Institutional Warrants in General So far, we have confined our discussion to the legal paradigm for institutional warrants. Can this discussion be generalized for the entire class of such warrants? The answer is straightforward. Institutions consist of declared rules. The legal case includes both constitutive and regulative rules. We have already seen how both legal constitutive rules and regulative rules can be analyzed as having the form ‘DP,’ where ‘D’ is the declarative operator and ‘P’ indicates some propositional content. Consider the constitutive rule Kicking the ball into the net counts as scoring a goal in soccer.

In the general form X counts as Y in context C

(Searle, 1969, p. 35), instances of ‘X’ are instances of types of action or behavior. But we cannot have an action without an actor (and where the action is transitive, an object of that action). So taking ‘A’ as a schematic letter for a subject-verb-object action, the form of a constitutive rule can be straightfowardly paraphrased as

146

J. B. Freeman

if Axy & Cxy, then Yxy if x kicks y into the goal and x is playing soccer with y, then x has scored a goal with y

The truth of this rule’s being a rule of soccer presupposes that it has been declared, by common consent if not by higher authority. Hence, the rule’s being in force presupposes the truth of D(Kicking the ball into the net counts as scoring a goal in soccer.)

As we have pointed out in Sect. 9.3, to understand this sentence, one must understand the rules of the declarative operator D and the meaning of the propositional content of the rule, a matter of knowing the semantic rules for the expressions involved. Nothing in understanding the rules of the declarative operator or the semantic meaning rules is specific just to understanding laws or the rules of a game. As long as the rules of an institution involve constitutive rules, our discussion of meaning in Sect. 9.3 applies. Now consider again Cohen’s paradigm of an elementary regulative rule schema: For any persons x and y, if Rxy, then if x sues y, x ought to win

(1970, p. 157). But clearly for “Rxy” to be a sufficient condition for “if x sues y, x ought to win,” “Rxy” must be declared a sufficient condition by an agent with authority. Hence, the rule’s being in force presupposes the truth of D(For any persons x and y, if Rxy, then if x sues y, x ought to win).

Hence understanding a regulative rule requires understanding the declarative operator and the propositional content of the rule. Such understanding involves nothing specific to laws or any particular institution. As we have seen in the legal case, understanding the propositional content of a rule involves being able to paraphrase the rule as a universally generalized conditional. Can the propositional content of an institutional rule always be paraphrased as a universally generalized conditional? Cases not open to such a paraphrase, should there be any, pose no problem. Since corresponding to a warrant is a generalized conditional, if the propositional content of a rule could not be paraphrased as a universally generalized conditional, it would not back any warrant. So for warrants backed from above, institutional warrants, the ‘DP’ analysis suffices. The examples of rules we have just considered are paraphrased by unqualified generalizations. Hence the warrants they back are conclusive, not defeasible. But we have seen that judgments made according to legal rules can be overturned in some circumstances. The warrant the rule backs is subject to rebuttals. Does this hold also of other institutions? Constitutive rules will not be qualified. As Searle puts it, constitutive rules “are almost tautological in character, for what the rule seems to offer is part of a definition” (1969, p. 34). If one has hit a baseball across the first base line in a game of baseball, one has hit a foul ball. This is the definition of a foul ball in baseball. A constitutive rule specifies that a certain act done in a certain context has a certain meaning. Performing the act in that context constitutes a sufficient condition for the act to have that meaning.

9 Assessing Connection Adequacy for Arguments …

147

By contrast, regulative rules are open to exception. For example, When an elderly person or one with disabilities or a pregnant woman gets on a bus, (and you are seated), give your seat to that person.

We may paraphrase the rule this way: If x is elderly or x is a person with disabilities or x is a pregnant woman and x gets on a bus y and there is no vacant seat on y and z is on y and z is seated, then z should give up z’s seat to x.

Here, the “should” is the should of behavior or proper behavior. This rule can certainly have exceptions. What if z him or herself is elderly, a person with disabilities, or a pregnant woman? Hence the warrant licencing the move from the antecedent as premise to the consequent as conclusion is subject to rebuttals. How strong is this warrant? One factor in determining strength will be the plausibility that one will oneself be elderly, a person with disabilities, or a pregnant woman. Consider this example from the institution of money: If x gives storekeeper y a $10 bill to pay for an item marked $8.99, then x has paid y in full for z.

Certainly an instance of the corresponding argument would give us strong reason for believing that x has paid y in full for z. But is the argument defeasible? Suppose the $10 bill is a counterfeit or that y is selling stolen property or that the item has been marked incorrectly, lower than the actual price? Under these conditions, x has not paid y in full for z or x’s doing so does not have its ordinary meaning. Applying the considerations on strength for defeasible argument backed by legal institutions which we have developed, we can view the rebutting conditions as relevant variables. How old or infirm must z be to be excused from giving up z’s seat to x? How advanced must z’s pregnancy be to likewise excuse z from giving up z’s seat? In the second example, the variables may not come in degrees but will either hold or not hold. A bill is either counterfeit or it is not. An item is stolen merchandise or it is not. The price tag is either correct or it is not. If we knew that one or more of these conditions held, the argument would be defeated. If it is just possible that one or more of them holds but we do not know whether one or more do hold, what does this mean for the strength of the argument? For each of the relevant variables, we may ask how plausible it is that the variable holds at least to some degree, where again plausibility depends on compatibility with previous knowledge. We may then rank the relevant variables according to plausibility, starting with the most plausible. Should we find out that the most plausible relevant variable does not hold, that counterrebutting information increases the strength of the argument. If the only relevant variables or the only remaining relevant variables all have low plausibility, then the argument may be rebuttal tolerant. Our adaptation of the legal case has been carried out without regard to particular institutions. It is a general approach to argument strength and argument cogency for arguments backed by institutional warrants. We rest our case. We have presented a method for assessing the strength of institutional warrants and for judging whether arguments with such warrants are connection adequate.

148

J. B. Freeman

References Cohen, L. J. (1970). The implications of induction. London: Methuen and Co., Ltd. Freeman, J. B. (1991). Dialectics and the macrostructure of arguments. Berlin and New York: Foris Publications. Freeman, J. B. (2005). Acceptable premises: An epistemic approach to an informal logic problem. Cambridge, MA: Cambridge University Press. Freeman, J. B. (2011). Argument structure: Representation and theory. Dordrecht: Springer. Freeman, J. B. (2013). What types of arguments are there? Studies in Logic, 6, 30–55. Freeman, J. B. (2018). Inferences, inference rules, generalized conditionals, adequate connections. In S. Oswald, & Maillat, D. (Eds.), Argumentation and Inference: Proceedings of the Second European Conference on Argumentation, Fribourg 2017. (pp. 79–100). London: College Publications. Searle, J. R. (1969). Speech acts: An essay in the philosophy of language. Cambridge, MA: Cambridge University Press. Searle, J. R. (1979). Expression and meaning: Studies in the theory of speech acts. Cambridge, MA: Cambridge University Press. Toulmin, S. (1958). The uses of argument. Cambridge, MA: Cambridge University Press.

Chapter 10

On the Logical Ways to Counter an Argument: A Typology and Some Theoretical Consequences Hubert Marraud

10.1 Introduction Elsewhere (Marraud, 2015) I have distinguished two conceptions of dialectic. Dialectic can be conceived of as the art of controversy or debate, with confrontation of opinions and hence of arguers. The focus of dialectics thus understood is the conventional rules and procedures governing such confrontations. This is what I call arguers’ dialectic. But dialectic can also mean the study of the oppositions between arguments. This arguments’ dialectic is historically linked to the notion of argument strength, so that arguments’ dialectic may be defined as the study of argument strength. The concept of counterargument lies at the heart of argument’s dialectic. The importance attached to the relations between arguments—especially those of opposition—establishes a clear distinction between the theory of argument and formal logic. My purpose in this paper is twofold. First, to give a suitable definition of counterargument; second, to establish a typology of counterarguments. I will confine myself to logical counterarguments: arguments criticizing another argument on account of its logical properties. Thus, ways of attacking an argument for being ineffective for a given audience or for procedural matters will not be taken into account. Let us begin with some preliminary definitions. To argue is to present to someone something as a reason for or against something else. A prima facie reason is something that appears to be (i.e., is presented as being) a reason, but may actually not be a reason at all. An argument gives a prima facie reason for (pro argument) or against (con argument) some claim. Hence a con argument gives a prima facie reason to reject some claim. These are more or less standard definitions, though many scholars H. Marraud (B) Department of Linguistics, Modern Languages, Logic and Philosophy of Science, Theory of Literature and Comparative Literature, School of Arts, Universidad Autónoma de Madrid, Madrid, Spain e-mail: [email protected] © Springer Nature Switzerland AG 2020 F. H. van Eemeren and B. Garssen (eds.), From Argument Schemes to Argumentative Relations in the Wild, Argumentation Library 35, https://doi.org/10.1007/978-3-030-28367-4_10

149

150

H. Marraud

define argument either in terms of its purported aim (justify, persuade, etc.) or of its structure (premises and conclusion). According to the Merriam-Webster Dictionary to counterargue is to give (reasons, statements, or facts) in opposition to an argument or in support of an opposing argument. Thus a counterargument is a reason to reject some argument. Therefore there is a clear difference between a con argument, a reason against a claim, and a counterargument, a reason against the cogency of an argument.

10.2 Main Kinds of Counterarguments To argue is to present something to somebody as a reason for (or against) another thing. Hence the minimal autonomous unit of argument is a compound of one or more premises expressing a reason, the “something” in the above definition, and a conclusion, the “another thing” in the same definition. When a reason for a conclusion is presented to someone but she does not endorse it, there are four possible moves. 1. She might reject the argument as a whole, alleging that there must be something wrong in it. 2. She might challenge some of the premises: “Where do you get that from?”. 3. She might question instead whether the reason offered is actually a reason for that conclusion: “What does that have to do with the conclusion?” 4. Finally she might try to show that the conclusion should still be rejected, offering an opposite reason: “Yes, but…”. Therefore there are four main strategies for logically attacking an argument: challenging the argument as a whole, challenging some of its premises, challenging the inferential step, or challenging its conclusion. But counterarguing requires more than challenging some of the parts of an argument: one has to give reasons for rejection. Thus one can either meta-argue that the offered argument is not cogent or argue that some of the premises are not true or acceptable, that the premises are not adequately connected to the conclusion, or that the conclusion is nonetheless false or unacceptable. Accordingly I shall distinguish four main kinds of counterargument, which I term, respectively, dismissal, objection, rebuttal and refutation. These four kinds are successively described below.

10.3 Dismissal An argument can be logically rejected as a whole for a variety of reasons; for instance, it can be adduced that argument A has been rejected by most philosophers, that A is not a scientific argument, that A leads to an unreasonable conclusion, etc. The

10 On the Logical Ways to Counter an Argument …

151

rationality of such holistic counterarguments has been advocated by Cohen (2001). Here is an example: […] as we saw before, states and corporations differ in one crucial respect. The shareholders of a corporation voluntarily take on the obligations of the corporation when they purchase shares; indeed, the corporation’s obligations are reflected as a discount in the price of a share. People who are born into citizenship of a state do not consent in a similar manner to take on the obligations that others have acquired in the name of the state. Although Locke argued that people give implicit consent to their government by not emigrating, no one takes this argument seriously anymore. Consent requires more than the ability to choose an extremely disagreeable alternative. (Posner, 2003: 1906).

For present purposes Locke’s argument can be summarized thus: A. People give implicit consent to their government by not emigrating, so people have an obligation to live up to the state’s obligations. The corresponding counterargument then runs as follows: CA. No one takes this argument seriously anymore. The implicit conclusion here is that Locke’s argument must be unsound. Thus a dismissal is an argument whose conclusion is an assertion to the effect that a given argument is not cogent. Since a meta-argument is “an argument about one or more arguments or about argumentation in general” (Finocchiaro, 2013: 1), dismissals are meta-arguments. Some fallacy charges can also be seen as dismissals of some kind, as shows the following example. A judge has granted a conditional discharge to a Eureka, Pictou Co. woman [Emily Anne MacIntosh] who stole more than $25,000 from the special care home where she worked as an administrator earning $99,000 annually. […] Her lawyer tried to argue the offence “is rendered somehow less serious as ‘there is no evidence of the government going insolvent’ or as ‘nobody missed out on lunch,’” [Pictou provincial court Judge Del] Atwood said. “In my view, this is a fallacy of false simplicity.” (Lambie, 2018).

The argument by the lawyer can be reconstructed as follows: A. There is no evidence of the government going insolvent due to MacIntosh’s action. So, MacIntosh’s offence is less serious than it appeared. This argument is then dismissed by Lambie as fallacious: CA. The argument that since there is no evidence of the government going insolvent, MacIntosh’s offence is less serious than it appeared is a fallacy of false simplicity, and so is not cogent.

10.4 Objection An argument A is an objection to an argument B when A’s conclusion is incompatible with some of the premises of B. This is obviously the case when they are mutually

152

H. Marraud

contrary or contradictory, but in argumentation there are also less stringent forms of incompatibility. Here is an illustration of this wider sense of incompatibility: A.

This patient has a streptococcal infection, so presumably this patient needs penicillin treatment. CA. Infection diagnosis is only based on symptoms, no clinical tests have been performed. The implication of the reply is not that the premise in the original argument is false, but that since no clinical tests have been performed, the speaker is not entitled to assert it. This can be diagrammed thus: Infection diagnosis is only based on symptoms and no clinical tests have been performed So It cannot be asserted that the patient has a against streptococcal infection

This patient has a streptococcal infection So This patient needs penicillin treatment

A successful objection suspends the conclusion.

10.5 Rebuttal I am using “rebuttal” in a sense close to Toulmin (2003). To prevent possible confusions, notice that Pollock (2007) uses “rebutting defeaters” and “undercutting defeaters” for something close to what I call “refutations” and “rebuttals”, respectively. The question “How do you get from premises to conclusion?” can be answered in either of two basic ways: (a) Resorting to an analogy: step from P to C is like (is analogous to) the already accepted step from P to C ; or (b) Providing a warrant, i.e. a general, hypothetical statement, which can act as a bridge, and authorizes the sort of step to which our particular argument commits us (Toulmin, 2003, p. 91).1 Mutatis mutandis, the inference proposed by an argument can be rejected by arguing either that it is analogous to another already rejected inference (counteranalogy or rebuttal by comparison), or that its warrant is problematic (warrant rebuttal). As for its effects, rebuttal is similar to objection in that a successful rebuttal suspends the conclusion. Let us illustrate first rebuttal by comparison (counteranalogy) with an example from Pies (2018). 1 See van Laar (2015) for a discussion of the different ways in which the adequacy of the connection

between the premises and the conclusion could be defended.

10 On the Logical Ways to Counter an Argument …

153

A.

As a physician and medical ethicist, I am opposed to any form of physician assistance with a patient’s suicide. The Hippocratic Oath—arguably, the most important foundational document in medical ethics—clearly states: “I will neither give a deadly drug to anybody if asked for it, nor will I make a suggestion to this effect.” CA. There was a time and place for the Hippocratic Oath to work, it doesn’t mean it’s always appropriate for all situations far into the future. That would be like the idiots who argue that we should always be allowed to keep guns just because it’s in the Constitution. The corresponding diagram looks like this: The fact that it’s in the Constitution doesn’t mean that we should always be allowed to keep guns Both arguments are analogous:

So

The fact that any form of physician assistance with a patient’s suicide is banned by the Hippocratic Oath doesn’t against mean that physicians and medical ethicists should oppose to it now:

The Hippocratic Oath clearly states: “I will neither give a deadly drug to anybody if asked for it, nor will I make a suggestion to this effect.” So Physicians and medical ethicists should oppose to any form of physician assistance with a patient’s suicide

A warrant can be held to be problematic by one of two reasons: either the warrant is generally invalid, so that conclusions like C cannot be inferred from premises like P through W (Plain rebuttal), or the warrant cannot be applied in the case at hand, so that C cannot be inferred from P through W (Exception). There are weaker forms of these (Disclaimers), since the validity or applicability of the warrant may be controversial, so that C can only be inferred with reservations from P, “at your own risk”. Thus we have four kinds of rebuttals: plain rebuttal, exception, warrant disclaimer and disclaimer by possible exception. • Plain rebuttal A.

Since you want to live comfortably, you must attend university, for university graduates usually do land jobs with healthy salaries. CA. The reason why university graduates usually do land jobs with healthy salaries is that they normally come from wealthier families and they are cleverer than the average.

154

H. Marraud

Or in diagrammatic form: The reason why university graduates usually do land jobs with healthy salaries is that they normally come from wealthier families and they are more clever than the average You want to live comfortably

So It is not true that university graduates usually do land jobs with healthy salaries just for being university graduates

University graduates against usually do land jobs with healthy salaries:

So You must attend university

• Exception A.

Polls give small advantage to candidate White over candidate Brown, so most likely White will win the election. CA. The advantage of candidate White over Brown is within the polls’ margin of error. Making explicit the warrant of the rebutted argument, we get the following diagram: The advantage of candidate White over candidate Brown is within the polls’ margin of error Polls give small advantage to candidate White over candidate Brown

So Polls cannot predict the winner of the election:

Against

Polls like this are generally reliable:

So White will most likely win the election

• Warrant disclaimer2 A.

Primed for a November 7th release and coming from British powerhouse stable Working Title, the new film from James Marsh is a likelihood for a fall festival bow. Its biopic structure and potential crowd pleaser appeal hints at Toronto rather than Venice. CA. Well, perhaps. But that’s a pretty unreliable rule of thumb, and changing all the time.

2 Taken

from “The Fall Festival 50,” (2014).

10 On the Logical Ways to Counter an Argument …

155

That is just a rule of thumb, and changing all the time

So

Biopics with a The rule that biopics with a potential crowd potential crowd pleaser appeal pleaser appeal are are more likely to be rewarded Against more likely to be in Toronto than in Venice is not rewarded in Toronto totally reliable than in Venice:

“The Theory of Everything” is a likelihood for a fall festival bow. It has a biopic structure and a potential crowd pleaser appeal

So

The Theory of Everything” is a likelihood for Toronto rather than for Venice

• Disclaimer by possible exception A.

With the falling of the pound, Germany, for instance, will be a more attractive place for Spanish, Portuguese, Polish and Greek emigrants than the UK. CA. Unless British companies raise wages to prevent the loss of labor force.

It could be the case that British companies raise wages to prevent the loss of labor force So

The stronger the UK could remain attractive for currency, the more UE immigrants despite the fall Against attractive the country of the Pound against the Euro for immigrants:

The Pound will fall against the Euro

So Germany will be a more attractive place for Spanish, Portuguese, Polish and Greek emigrants than the UK

10.6 Refutation As I stated in section 1 a prima facie reason is something that appears to be (i.e. is presented as being) a reason, but may actually not be a reason at all. It follows from the preceding definition of arguing that any argument offers a prima facie reason for its conclusion. Objecting to and rebutting are different ways of arguing that the alleged reason is not actually a reason. Prima facie reasons should not be confused with pro tanto reasons. A pro tanto reason is a consideration that counts in favor of

156

H. Marraud

some position, and thus has genuine weight, but nonetheless may be outweighed by other considerations (Kagan, 1989, p. 17n). A refutation is an attempt to show that a pro tanto reason is outweighed by other consideration, so that it is inconclusive. Thus it is part of the pragmatics of refutation that its target be an argument that is taken to present a pro tanto reason. I distinguish three kinds of refutation, associated with the use and meaning of these three the phrases: A but B, A but also B, and Although B, A. In many cases when someone utters A but B she means that (1) she accepts A, (2) she accepts B, (3) A is a reason for some conclusion C, (4) B is a reason for some conclusion C  incompatible with C, and (5) in the situation of utterance, B outweighs A. When A but B is used in this way, the addressee is invited to infer C  from the joint consideration of A and B. In these cases I will say that the argument B, so C  is a contradicting refutation of the argument A, so C. A.

This patient has a streptococcal infection, so presumably this patient needs penicillin treatment. CA. But this patient is allergic to penicillin. Conflating both arguments into a single argumentation, we get: This patient has a streptococcal infection So This patient needs penicillin treatment

but
It is impossible not to refer briefly to the Brahimi report and some of its main recommendations

The Brahimi report does not answer all of our questions So It is not necessary to refer to the Brahimi report in every circumstance

10.7 Weighing Any refutation involves the weighing of pros and cons, since the conclusion is reached from the joint examination of positive and negative considerations. This feature relates refutation to Wellman’s third pattern of conduction. The third pattern of conduction is that form of argument in which some conclusion is drawn from both positive and negative considerations. In this pattern reasons against the conclusion are included as well as reasons for it. For example “in spite of a certain dissonance, that piece of music is beautiful because of its dynamic quality and its final resolution” or “although your lawn needs cutting, you ought to take your son to the movies because the picture is ideal for children and will be gone by tomorrow.” (Wellman, 1971: 57).

Weighing is characteristic of refutation as a distinctive kind of counterargument. While objection and rebuttal look for flaws or faults in the argument, refutation starts from the acceptance of the argument under examination, as the analysis of argumentative connectors like “but” and “although” shows. There is nothing wrong on a refuted argument except that there is a stronger argument to the contrary. The distinction of different kinds of counterarguments is important to limit the scope of weighing. Some authors seem to consider weighing as an all pervasive phenomenon in argumentation. Robert Pinto, for example, writes: “In short, we can identify ‘exceptions’ to a qualified generalization only if we are already able to compare the strength of arguments licensed by that generalization to certain other arguments.” (2011: 117). Although this issue deserves a more detailed discussion,

158

H. Marraud

the unbounded spreading of weighing promoted by such argumentation theorist as Pinto is the result of mixing together the degree of acceptability of the premises and the robustness or strength of the link between premises and conclusion. To borrow an example overused in AI, there is no point in comparing (if it is possible at all) the strength of the argument Tweety is a bird therefore Tweety flies since birds typically fly with the strength of the rebutting argument Tweety was watched in Peninsula Valdés, where there are important colonies of penguins, so probably it is a penguin. In order to assess the later the sensible thing is to compare its strength with that of Tweety was watched in Punta Loma, so it is more likely a rock shag than a penguin. Of course the possibility exists that the force of two arguments with compatible premises be incomparable. So far as arguments are concerned, force or strength defines in the best case a partial order. It might seem that whenever two such arguments are conjoined it should be prudent to suspend judgment on the issue at stake. This skeptical position can be contrasted with a credulous one, according to which both conclusions are tenable, and in order to decide which argument is in force in some particular argumentative situation one has to resort to procedural rules. I take the terms “skeptical” and credulous” from related discussions in IA (Prakken & Vreeswijk, 2002, Sect. 4.1).

10.8 Rebuttals, Refutations and Standards of Proof It might be thought that sufficiency is a comparative notion to be explained in terms of weighing. However, sometimes the stronger argument can be rejected for failure to comply with the standards of proof of the given dialogue. These standards of proof have to do with acceptable risks and vary from one type of dialogue to the other. Although perhaps this could be described as a situation in which the argument has not the required weight, such cases cannot be classified as refutations for no comparison of the strength of two opposite arguments is involved. Here is a putative example (drawn from Collinson, 2007): A.

The wingbeat frequency of the bird in the Luneau video was measured at 8.6 beats s−1 , similar to that inferred from archival sound recording of a single Ivory-billed Woodpecker, but claimed to be outside the range of Pileated Woodpeckers. So the woodpecker in flight recorded by Luneau is an Ivorybilled Woodpecker. CA. However, to regard the Luneau video by itself as presenting anything other than an unidentified woodpecker falls below the standards of proof normally required for scientific publication: the images are not good enough. This is not a contradicting refutation since no reason is being offered to believe that the woodpecker in flight recorded by Luneau is not an Ivory-billed Woodpecker. The point is rather that given that the images are not good enough, they don’t provide, according to the current standards of proof for scientific publications, a genuine reason to accept that the woodpecker in flight is an Ivory-billed Woodpecker. So the

10 On the Logical Ways to Counter an Argument …

159

Table 10.1 Orientation of counterargument structures Argument

Counterargument

Joint orientation towards C

A so C

Objection

Neutral

Rebuttal by comparison

Neutral

Plain rebuttal

Neutral

Exception

Neutral

Disclaimer

Positive

Contradicting refutation

Negative

Cancelling refutation

Neutral

Diminisher

Positive

allegation that an argument fails to comply with the current standards of proof counts as a case of rebuttal.

10.9 Counterargument Structures The previous discussion brings to light that counterargumentation produces complex argument structures, depicted by the previous diagrams. An argument structure can be defined as a combination of arguments with an explicit argumentative orientation (positive, negative or neutral) towards some claim. The point is that when one counterargues one takes charge of the criticized argument, and therefore combines the criticized and the critical argument into a single argumentation. The join consideration of an argument A so C and an objection is neutrally oriented towards C,3 as does its joint consideration with a plain rebuttal or a cancelling refutation (marked by “but also”). By contrast, the conjoining of A so C with a contradicting refutation is negatively oriented towards C, while its conjoining with a diminisher is positively oriented towards C (Table 10.1). This idea takes us directly from arguers’ dialectic to arguments’ dialectic. Arguers’ dialectic is dynamic while arguments’ dialectic is static; static diagrams are to arguments’ dialectic what dynamic profiles of dialogue are to arguers’ dialectic4 . Arguments’ dialectic is part of the theory of argument, and given that logic is concerned with arguments as products, it is part of logic. Freeman argues that where a logical analysis takes as its unit of analysis the individual argument, a dialectical analysis of argumentation focuses on the whole argumentation, possibly containing multiple arguments that need not form one single unified argument developed over the course 3 An

objection to A, so C is neutrally oriented towards C insofar as it neutralizes the positive orientation of A towards C. In such a case neither C nor not C can be concluded on this basis alone. 4 “A profile of dialogue is a connected sequence of moves and countermoves in a conversational exchange of a type that is goal-directed and can be represented in a normative model of dialogue.” (Walton, 1999: 53).

160

H. Marraud

of the discussion. (2011: 109). But taking in account the way counterarguments logically relate to their target argument, it is also possible to analyze such a relation as an additional kind of (counter)argumentative structure. It is like seeing just one frame of a movie. Accordingly, my contention is that the standard catalogue of argument structures, which includes subordinate, multiple and coordinate argumentation, needs to be supplemented with counterargument structures: objection, rebuttal and refutation. Moreover, I claim that there is a close relation between both kinds of structures. A natural reply to an objection is to ground the contested premise offering some reason for its acceptance. The product of this move is a chain of arguments, a subordinate argumentation, whose first link works as a contradicting refutation of the alleged objection. Backing is a plausible reaction when we try to defend an argument from a plain rebuttal. In Toulmin’s model “When I challenges A to produce ‘further substantial supporting considerations’—and so demonstrate that his warrant is sound and relevant—I requires A to provide that warrant with backing.” (Toulmin, Rieke, & Janik, 1984: 63). Let us resort to a well-known example from Toulmin; imagine someone arguing this way:

A Swede is almost certainly not a Roman Catholic:

Petersen is a Swede So Petersen is almost certainly not a Roman Catholic

Then the opponent tries to rebut this argument pointing out that the number of Roman Catholics is growing steadily in Sweden: The number of Roman Catholics is growing steadily in Sweden So It is no longer certain that a A Swede is almost Swede is almost certainly against certainly not a Roman not a Roman Catholic Catholic:

Petersen is a Swede So Petersen is almost certainly not a Roman Catholic

To defend her argument the proponent can produce further substantial support to demonstrate that her warrant is reliable, e.g. as of 2015, legally registered Catholics in Sweden were 113,053, comprising 1.2% of the total population. Adding this backing into our diagram we obtain:

10 On the Logical Ways to Counter an Argument …

161

The number of Roman In 2015 legally registered Catholics in Catholics is growing but Sweden were 113,053, comprising steadily in Sweden 1.2% of the total population So < So Petersen is a Swede It is no longer certain that a A Swede is almost certainly not a So Swede is almost certainly against Roman Catholic: not a Roman Catholic Petersen is almost certainly not a Roman Catholic

Thus backing the warrant results in an argument (In 2015 legally registered Catholics in Sweden were 113,053, comprising 1.2% of the total population; so a Swede is almost certainly not a Roman Catholic) acting as a refutation of the previous rebuttal (The number of Roman Catholics is growing steadily in Sweden; so it is no longer certain that a Swede is almost certainly not a Roman Catholic). Finally, a coordinative argumentation is a combination of mutually reinforcing reasons which only taken together provide sufficient support for their common conclusion (van Eemeren & Grootendorst, 1992: 73–89; Snoeck-Henkemans, 2000). As we have seen, a refutation of a given argument consists in an argument with an opposite conclusion that is presented as being stronger. The arguer can attempt to meet this refutation by adding a new reason to produce a stronger conjunction of reasons. The resulting coordinative complex argument will act, once again, as a refutation of the former refutation.

10.10 Other Classifications Blair and Johnson (1987) propose a classification of counterarguments based on the RSA (Relevance, Sufficiency, Adequacy) criteria of good argument: an argument is cogent if and only if it has acceptable premises, its premises are relevant to its conclusion, and the premises provide sufficient grounds for the conclusion. Hence it can be argued that an argument is not cogent for one of two reasons: either its premises are unacceptable, or the move from its premises to its conclusion is problematic. In turn, the step from premises to conclusion can be problematic either because the alleged reason is irrelevant (i.e., it is not really a reason at all), or because it is an insufficient reason. This means that Blair & Johnson’s theory make no room for refutation, even if this is a quite common strategy for attacking an argument. An examination of counter-argumentative procedures in polemic dialogues leads Quiroz, Apothéloz & Brandt (1992) to propose a quadripartite classification. (1) Counterarguments concerning the plausibility of the reason: they challenge the plausibility of some of the premises of the argument at issue; these correspond to my objections.

162

H. Marraud

Table 10.2 Walton classification of counterarguments Objections

Procedural

Requests for

Objections

Clarifications

Challenges

Irrelevance

Rebuttals

Non -rebutting challenges

Internal

Refutations

Rebuttals

External

Exceptions

Rebuttals

Premise

Internal

External

Attacks

Refutations

Refutations

(2) Counterarguments concerning the completeness of the reason: they offer a stronger reason for a conclusion incompatible with the conclusion supported by the contested reason; these are what I have called “refutations”. (3) Counterarguments concerning the relevance of the reason: they dispute the relevance of the premises to the conclusion; thus these counterarguments amount to plain rebuttals. (4) Counterarguments concerning the argumentative orientation of the reason: they contend that the reason given is really a reason for a different, opposite conclusion; these seem to be a special sort of refutations. Walton (2013: 27–62) offers a more complex (though not necessarily more clarifying classification) (Table 10.2). Walton defines an objection as any way of questioning the use of an argument on a given occasion. However, in fact, he only considers dialectical (procedural) and logical objections (challenges), letting aside rhetorical objections referred to audience reception (save perhaps requests for clarifications). A challenge is an expression of critical doubt about whether a reason supports the argument. Challenges are divided into rebuttals and exceptions. A rebuttal is an argument that is directed against other prior argument, in order to show that it is open to doubt or not acceptable. A refutation is a rebuttal that is successful in carrying out its aim. Now rebuttals are divided into external and internal ones. Internal rebuttals aim at the premises of the argument under scrutiny. There are two kinds of external rebuttals. First, an external rebuttal can be an argument to the effect that the conclusion doesn’t follow from the set of premises that were presented

10 On the Logical Ways to Counter an Argument …

163

as supporting it. Second, an external rebuttal can be an argument that is stronger than the argument that is challenged and that provides a reason for rejecting its conclusion. Walton defines an exception as a distinctive kind of premise, brought to light by means of critical questions, and classifies Pollock-style undercutters as exceptions. Pollock (2007) makes a distinction between rebutting defeaters and undercutting defeaters. A rebutter of an argument gives a reason to show its conclusion is false, whereas an undercutter merely raises doubt whether the inference supporting the conclusion holds.

10.11 The Order of Counterargumentation In real argumentative practices argument criticism is seemingly conducted according to a procedure that seeks to get the maximum effectiveness at the lowest cognitive cost. For this reason it can be observed that priority is given to objection over rebuttal, and then to rebuttal over refutation. Refutation can result in a costly process of weighing pros and cons, the outcome of which is uncertain. Leibniz said that the possession of a balance of reasons would be an even more important achievement than the “fabulous science of producing gold”. To understand why rebuttal is a strategy more onerous than refutation, one just has to think about the different functions of premises and warrants. Premises are presented as data or facts while warrants are rules. Accordingly premises are appraised as true or false and warrants as more or less reliable. Toulmin (2003: 98) puts it as follow: a general statement like “Philosophers in their intellectual plenitude are unmarried” (according to Pierre Riffard (2004), 70% of philosophers were unmarried when they published their masterpieces) can be used in two different ways: – as a datum, and then it can be expressed in the form “It has been observed that philosophers were unmarried when they were in their intellectual plenitude”; – as a warrant, and then it can be paraphrased “If a philosopher is in his intellectual plenitude, he may be presumed to be unmarried”. To demonstrate the falsity of a universal statement is easier than to demonstrate that it is not a reliable guide to making inferences, if only because a false statement is false in any circumstance, while a guide to inference can be reliable in some circumstances and unreliable in others. A pragmatic effect that confirms the effective operation of this cost-saving procedure ordering counterargumentation is that when someone attempts in the first place to rebut or to refute an argument, this can be taken as a sign of her acceptance of the premises. A pragmatic effect that confirms what has been said is that if someone attempts in the first place to rebut or to refute an argument, this can be taken as a sign of her acceptance of the premises.

164

H. Marraud

10.12 Conclusions A counterargument is an argument presenting something as a reason to conclude that some other argument is unsound, invalid, or not cogent (these are all terms of logical appraisal). A counterargument, a reason against an argument, should not be confused with a con argument, a reason against a claim. The most immediate reasons to reject an argument from a logical point of view are that either it does not present a prima facie reason or that it presents a pro tanto reason that is outweighed by other opposite reasons (refutation). An argument fails to present a prima facie reason when either it does not present a real fact (objection), or when the alleged fact is irrelevant (rebuttal). An argument structure is a combination of arguments with an explicit argumentative orientation, positive, negative or neutral. Counterarguing involves conjoining the criticized and the critical argument into a single argumentation. Therefore, counterargumentation produces argument structures that can be represented by diagrams. Thus the standard catalogue of argument structures, which includes subordinate, multiple and coordinate argumentation, needs to be supplemented with counterargument structures: objection, rebuttal and refutation. It further follows that the proposed classification of counterarguments is also a classification of argument structures (Table 10.3). Acknowledgements This work has been made possible by funds provided by the Spanish Ministry of Economy and Competitiveness through the Research Project FFI2014-53164-P, “The construction of argumentative agents in the practices of public discourse”.

Table 10.3 Proposed classification of counterarguments Counterargument

Dismissal

Objection

Rebuttal

Rebuttal by comparison

Plain Rebuttal

Refutation

Warrant Rebuttal

Exception

Contradicting refutation

Warrant disclaimer

Cancelling refutation

Disclaimer by exception

Diminisher

10 On the Logical Ways to Counter an Argument …

165

References Blair, J. A., & Johnson, R. H. (1987). Argumentation as dialectical. Argumentation, 1, 41–56. Cohen, D. H. (2001). Evaluating arguments and making meta-arguments. Informal Logic, 21(2), 73–84. Collinson, J. M. (2007). Video analysis of the escape flight of Pileated Woodpecker Dryocopus pileatus: does the Ivory-billed Woodpecker Campephilus principalis persist in continental North America? PMC. Retrieved February 3, 2019 from https://www.ncbi.nlm.nih.gov/pmc/articles/ PMC1838407/. Finocchiaro, M. (2013). Meta-argumentation. An approach to logic and argumentation theory. London: College Publications. Freeman, J. B. (2011). Argument structure. Representation and theory. Dordrecht Heidelberg London New York: Springer. Lambie, C. (2018). Judge grants discharge to Pictou County woman in theft from special care home. The News. 22/11/2018. Retrieved February 7, 2019 from https://www.ngnews.ca/news/local/ judge-grants-discharge-to-pictou-county-woman-in-theft-from-special-care-home-261878/. Marraud, H. (2015). Do arguers dream of logical standards? Arguers’ dialectic vs. arguments’ dialectic. Revista Iberoamericana de Argumentación, num.10. Pies, R. W. (2018). How assisted suicide affects physicians. Undark Magazine 01/09/2018. Retrieved February 7, 2019 from https://undark.org/article/assisted-suicide-physician-health/. Pinto, R. C. (2011). Weighing evidence in the context of conductive reasoning. In J. A. Blair & R. H. Johnson (Eds.), Conductive argument: An overlooked type of defeasible reasoning (pp. 104–126). London: College Publications. Pollock, J. L. (2007). Defeasible reasoning and degrees of justification. Argument & Computation, 1(1), 7–22. Pollock, J. L. (2010). Defeasible reasoning and degrees of justification. Argument & Computation, 1(1), 7–22. Posner, E. A. (2003). Do States have a moral obligation to obey international law? Stanford Law Review, 1906, 55(5), 1901–1919. Prakken, H., & Vreeswijk, G. (2002). Logics for defeasible reasoning. In D. M. Gabbay & F. Guenthner (Eds.), Handbook of philosophical logic (2nd ed., Vol. 4, pp. 219–318). Dordrecht: Kluwer Academic Publishers. Quiroz, G., Apothéloz, D., & Brandt, P.-Y. (1992). How counterargumentation works. In F. H. van Eemeren, et al. (Eds.), Argumentation illuminated (pp. 172–177). Amsterdam: SicSat. Riffard, P. A. (2004). Les philosophes: vie intime. Paris: Presses Universitaires de France (PUF). Shelly, K. (1989). The limits of morality. Oxford: Clarendon Press. Snoeck-Henkemans, A. F. (2000). State-of-the-art: The structure of argumentation. Argumentation, 14, 447–473. The Fall Festival 50: Our Wishlist for The Venice, Telluride And Toronto Film Festivals. IndieWire 30/06/2014. Retrieved February 7, 2019 from https://www.indiewire.com/2014/06/ the-fall-festival-50-our-wishlist-for-the-venice-telluride-and-toronto-film-festivals-84321/. Toulmin, S. E. (2003). The uses of argument. New York: Cambridge University Press. Toulmin, S. E., Rieke, R., & Janik, A. (1984). An introduction to reasoning (2nd ed.). New York: McMillan. Van Eemeren, F. H., & Grootendorst, R. (1992). Argumentation, communication, and fallacies. A pragma-dialectical perspective. Hillsdale, NJ: Erlbaum. van Laar, J.-A. (2015). Connection premises: Their character, criticism, and defence. In C. Ilie & G. Garzone (Eds.), Argumentation in real and virtual environments: Cross-disciplinary perspectives (pp. 39–55). Amsterdam/Philadelphia: John Benjamins.

166

H. Marraud

Walton, D. N. (1999). Profiles of dialogue for evaluating arguments from ignorance. Argumentation, 13, 53–71. Walton, D. N. (2013). Methods of argumentation. New York: Cambridge University Press. Wellman, C. (1971). Challenge and response: Justification in ethics. Carbondale, IL: Southern Illinois University Press.

Chapter 11

Arguing for Questions David Hitchcock

11.1 Introduction In a predecessor to this paper (Hitchcock, 2019), I argued on the basis of 17 examples 1 that people sometimes argue for questions. In each example, after making one or more statements or asking a question, the author introduced an interrogative sentence with a combination of a conclusion indicator and an interrogative particle: so where, so why, so what, how then, so are, so which, so how, where then, so have, so when, therefore why. In 10 cases, the reason for posing the question was to provoke interest in the subsequent revelation of the correct answer; a typical example was the newspaper headline: “Your smart phone is making you stupid, antisocial and 2 unhealthy. So why can’t you put it down?” In six cases, the reason was to put one or more addressees, or a third party, on the spot—to challenge them to explain or justify their behaviour; a typical example was another newspaper headline: “Mayor 3 John Tory has got plans so when is the time for action?” In the remaining case, the author wrote: “What value does conciousness [sic] add, and therefore why did it ever

1 One of the examples came from a call to a radio station’s phone-in program; the others were found on the Web. 2 Eric Andrew-Gee, The Globe & Mail, 2018 01 06; available at https://www.theglobeandmail.com/ technology/your-smartphone-is-making-you-stupid/article37511900/; accessed 2019 01 09. 3 Edward Keenan, The Star, 2017 12 19; available at https://www.thestar.com/opinion/starcolumnists/2017/12/19/mayor-john-tory-has-got-plans-so-when-is-the-time-for-action.html; accessed 2019 01 09.

D. Hitchcock (B) McMaster University, Hamilton, Canada e-mail: [email protected] © Springer Nature Switzerland AG 2020 F. H. van Eemeren and B. Garssen (eds.), From Argument Schemes to Argumentative Relations in the Wild, Argumentation Library 35, https://doi.org/10.1007/978-3-030-28367-4_11

167

168

D. Hitchcock

come into existence?”4 The reason for posing the question was to solicit answers; not coincidentally, this was the only case whose concluding question was preceded by another question rather than by one or more statements.

11.2 Arguments for Questions? What are we to make of such discourses and texts? Are they really arguments, or do their apparent conclusion indicators mean something other than what they seem to mean? Are the authors of the concluding interrogative sentences really asking a question, or are they doing something else? The linking particles in these examples do seem to be functioning as conclusion indicators. That your smart phone is making you stupid, antisocial and unhealthy provides a reason for asking why you can’t put it down. That a city’s mayor has plans is a reason for asking when is the time for action. And an answer to the question what value consciousness adds helps to answer the question why consciousness came into being. Thus the words “so” and “therefore” in these examples indicate an inferential transition; they do not signify, for example, either a cause-effect relationship or a narrative flow. The other examples manifest a similar rational connection between what precedes the linking particle and what follows it. Hence such discourses and texts are arguments. It is not obvious from their interrogative form, however, that the conclusions of such arguments are really questions. The noun “question” has three distinct senses: a syntactic sense, a pragmatic sense, and a semantic sense (Groenendijk & Stokhof, 1996). In the syntactic sense, a question is a kind of sentence (an “interrogative sentence”) distinguished by various devices from other kinds of sentences, such as declarative and imperative sentences. In the pragmatic sense, a question is the illocutionary act typically performed in uttering an interrogative sentence; it is the “interrogative act” of asking for certain information. In the semantic sense, a question is the “thing” that is asked (the “question” in the narrow sense) when someone performs an interrogative act, i.e. the content of this act. To determine whether an interrogative sentence introduced by a conclusion indicator is really a question, we need to settle the ontological status of the conclusions of arguments. If the conclusions of arguments are propositions and the like, then the issue is whether an interrogative sentence in the conclusion slot expresses a question in the semantic sense. If on the other hand the conclusions of arguments are assertions and the like, then the issue is whether the author of an interrogative sentence in the conclusion slot is posing a question, in the pragmatic sense of asking for certain information.

4 The

question was posted in September 2012 on Quora, a website on which anyone can submit a question and invite answers from volunteers. It is available at https://www.quora.com/What-valuedoes-conciousness-add-and-therefore-why-did-it-ever-come-into-existence; accessed 2019 01 09.

11 Arguing for Questions

169

My own view, expressed and defended most recently in (Hitchcock, 2018), is that the premisses and conclusions of arguments are illocutionary act-types, and that the conclusion of an argument can be an illocutionary act-type of any of the five main kinds recognized by Searle (1976): a representative, a commissive, a directive, an expressive, or a declarative. On this conception, the issue concerning arguments with an interrogative sentence in the conclusion slot is whether the utterance or inscription of this sentence is an interrogative act (which is a kind of directive). Is the author really asking something? Not all utterances and inscriptions of interrogative sentences are interrogative acts that ask a question.5 In particular, so-called “rhetorical questions” are not interrogative acts and do not ask a question. A rhetorical question is an indirect quasi-assertion whose author illegitimately shifts the burden of proof by implicitly challenging the addressee to give a reason for disagreeing with the expected answer (Kraus, 2006; cf. van Eemeren, 1987). In the 17 examples with an interrogative sentence in the conclusion slot, it seemed clear that the author was asking a question. The author of the headline about smart phones was really asking why you cannot put your smart phone down, and was inviting you (the reader) to ask yourself the same question. The author of the headline about the mayor’s plans was really asking when there would be action on those plans, and was inviting the reader to ask the mayor and council the same question. The person who submitted the question to Quora really wanted to find out what value consciousness added, as a way of answering the question why it came into existence. The other examples similarly concluded with a genuine question. They are not easily construed as rhetorical questions. In most of them, in fact, the concluding interrogative sentence was a “why” or “how” question, to which there is no obvious or expected answer. In the other cases, the interrogative sentence was clearly information-seeking: Which type of drunk are you? What kind of oil should you cook with? Which property development companies’ stocks are likely to do well in the next year? Again, there was no obvious or expected answer that would allow one to construe the interrogative sentence as a rhetorical question. Thus there are arguments for questions. Advancing them seems to have the immediate purpose of putting the question on the table. Arguing for a question should be understood as an attempt to provide a basis for asking it, i.e. as an attempt to show that the question needs to be answered. It is the analogue in interpersonal communication of the fact that the questions that we ask ourselves do not just come out of the blue; they arise from recognition of some perplexing phenomenon or event. It is of course possible to interpret what appears to be an argument for a question as really an argument for something else. Such interpretations are necessary as a defence of comprehensive theories of argumentation that require the conclusions of arguments to be of a single type. We can consider two examples of such theories. The pragma-dialectical approach of van Eemeren and Grootendorst (1984, 1992, 2004) construes argumentation as a complex of assertives put forward in an attempt to 5 Nor do speakers and writers who ask a question always do so by uttering an interrogative sentence.

One can ask a question, for example, by means of an imperative sentence like “Tell me how you are feeling”.

170

D. Hitchcock

justify or refute a proposition, in the context of a critical discussion whose function is to resolve an expressed difference of opinion about the proposition. Within the pragma-dialectical approach, an argument for a question can be interpreted as an attempt to resolve a difference of opinion about the proposition that the question is worth investigating. The epistemological approach of Lumer (1990, 1991) holds that the conclusion of an argument must be a judgment that a definite proposition is true or acceptable, on the ground that an argument is an instrument whose standard function is to produce knowledge in an addressee who knows the premisses but not the conclusion. Within this epistemological approach, an argument for a question could be interpreted as an attempt to produce knowledge in the addressee that the question is worth investigating. Aside from such antecedent theoretical commitments, however, there is no reason to construe arguments for questions as anything other than what they appear to be: arguments for asking a question. And, since arguing for something is trying to justify it, there are attempts to justify acts of asking a question. To justify something is to show that it is correct. Hence, an argument for a question is an attempt to show that it is correct to ask it. Correctness here does not mean truth or warranted acceptability or any such truth-like status. It means a combination of soundness (in the sense of the existence of a correct direct answer to the question asked) with ignorance of what that answer is.

11.3 The Formal Logic of Inferences to Questions What is the logic of arguments for questions? Under what circumstances does the set of statements offered in support of a question justify asking it? How can asking one question justify asking another question? To address these questions, we can consult the “inferential erotetic logic” of the Polish logician Wi´sniewski (1996, 2013)—a logic of “erotetic inferences”, i.e. inferences to questions. Wi´sniewski’s basic insight in developing this logic is that interrogative acts do not just occur; they arise. In thought or speech, they arise in two different ways. In an “erotetic inference of the first kind” they arise from beliefs, or, to put it linguistically, from assertions. In an “erotetic inference of the second kind” they arise from another interrogative act, perhaps with the help of one or more beliefs or assertions. For each kind of inference, Wi´sniewski proposes an account of when the inference is valid, i.e. when the question in some sense “follows from” the premisses. Wi´sniewski’s account applies to any formal language with a dichotomy between declarative sentences and interrogative sentences (which in a formal language he calls “erotetic formulas”). In such a language each declarative sentence has a truthvalue, once its non-logical symbols are assigned an interpretation. The logic for these sentences can be anything one likes: classical or intuitionist or relevantist, propositional or first-order or modal, and so on. The erotetic formulas consist of a question mark followed by a sequence of two or more declarative sentences within braces and separated by commas. They represent the logical structure (at a certain

11 Arguing for Questions

171

level of abstraction) of a natural-language interrogative sentence with a known finite set of direct answers. For example, the interrogative sentence “What kind of oil should you cook with?” has a finite set of direct answers: canola oil, coconut oil, corn oil, cottonseed oil, olive oil, palm oil, peanut oil, rapeseed oil, safflower oil, sesame oil, soybean oil, sunflower oil.6 We can construct a formal language in which the English names of the oils represent the English sentences of the form “you should cook with x”. In this formal language, the question “What kind of oil should you cook with?” would be symbolized as follows: (1) ?{canola, coconut, corn, cottonseed, olive, palm, peanut, rapeseed, safflower, sesame, soybean, sunflower}

Thus an erotetic formula is a question-inflected list of the direct answers to the question that it expresses. In other words, it represents an interrogative sentence as the set of its direct answers. An “erotetic inference of the first kind” is an inference from a set of declarative sentences to an erotetic formula. What does it mean for such an inference to be valid, i.e. for the erotetic formula to arise from the declarative sentences? As an analogy, consider the standard model-theoretic conception of the validity of an inference in a formal language from a set of declarative sentences to a declarative sentence: Validity of this sort is the absence of a counter-interpretation, i.e. the absence of an interpretation (of the inference’s non-logical symbols) on which the premisses are true but the conclusion is untrue. Wi´sniewski, sensibly enough, does not regard erotetic formulas as having truth-values. Rather, they have what we might call a “soundness-value”: on a given interpretation of their non-logical symbols, they are sound if they have a true direct answer and otherwise unsound. For an erotetic formula to arise from a set of declarative sentences, the set must entail (i.e. logically imply) that the erotetic formula is sound in this sense. That is, on any interpretation on which the declarative sentences are true, the erotetic formula has a true direct answer. Wi´sniewski (2013, p. 51) calls this condition “transmission of truth into soundness”. Transmission of truth into soundness is however not enough to make an inference to an erotetic formula valid, because this condition is met by any inference to an erotetic formula from one of its direct answers. For example, it would be met by an inference to our erotetic formula “?{canola, coconut, corn, cottonseed, olive, palm, peanut, rapeseed, safflower, sesame, soybean, sunflower}” from the sentence “canola”, an inference that would represent the natural-language inference to the question “what kind of oil should you cook with?” from the assertion “you should you cook with canola oil”. But validity of an erotetic inference of the first kind is supposed to capture the informal notion of a question arising from given information. If the given information includes or entails a direct answer to the question, then the question does not arise; it has already been answered. Thus the second condition for validity of an erotetic inference of the first kind is that the premisses do not entail a 6 For

simplicity, I restrict the list to the major oils in Wikipedia’s list of vegetable oils (https:// en.wikipedia.org/wiki/List_of_vegetable_oils; accessed 2019 01 09). It would be longer if all the cooking oils mentioned in that article were listed, and longer still if disjunctive answers (e.g. “canola or sunflower oil”) were included.

172

D. Hitchcock

direct answer to the erotetic formula, a condition that Wi´sniewski (2013, p. 51) calls “informativeness”. He uses the term “evocation” for the relation between premisses and conclusion in an inference that meets the two conditions of transmission of truth into soundness and informativeness. A set of declarative sentences evokes an erotetic formula if and only if the set entails that the erotetic formula has a true direct answer but does not entail any particular direct answer. The concept of evocation does not require that the premisses that evoke an erotetic formula are actually true. It is a conditional concept: If the premisses are true, then the conclusion has at least one true direct answer but the premisses do not settle which. Further, evocation is always general. The relationship of entailment is a function of the semantics of the logical symbols of the formal language and of the structure of the declarative sentences in the premisses and conclusion. The entailment of the conclusion’s soundness and the non-entailment of any direct answer are thus independent of the meaning and reference of the non-logical symbols of the language. We can illustrate this generality by representing formally the inference in the following example: (2) Scientists say there are four types of drunks – so which are you?7

For simplicity, let us abstract from the frame “scientists say” and take the premiss to be: that there are four types of drunks. This premiss could be represented in a first-order language with a symbol for exclusive disjunction, as follows: EITHER type 1(x) OR type 2(x) OR type 3(x) OR type 4(x).8

And the conclusion could be represented as follows: ?{type 1(you), type 2(you), type 3(you), type 4(you)}9

The inference from premiss to conclusion can be interpreted for any domain (or universe of discourse) over which the implicit universal quantifier ranges, any four one-place predicates defined on that domain, and any individual in that domain. For example, the domain could be taken to be giraffes, the four types could be taken to be the recently distinguished four species of giraffes (Fennessy et al., 2016), and “you” could be taken to be another name of a baby giraffe named “Tajiri” born in April 2017 in the Animal Adventure Park in upstate New York. Then the inference in our formal language would represent the following natural-language inference: (3) There are four species of giraffes – so which is Tajiri?

7 This

is the title of an article by Jasper Hamill, Metro, 23 Nov 2017 (available online at http://metro.co.uk/2017/11/23/scientists-say-there-are-four-types-of-drunks-so-which-are-you7103363/; accessed 2019 01 09). 8 An initial variable-binding universal quantifier “FOR EVERY x” is implicit in the use of the letter “x”, which by algebraic convention is a variable rather than a constant. 9 The formalization takes the liberty of treating the word “you” in the context as a proper name of whoever is reading the passage.

11 Arguing for Questions

173

The formal representation of examples 2 and 3 shows how Wi´sniewski’s concept of evocation can be used to evaluate natural-language inferences to interrogative acts. There must be a defensible formalization of the natural-language inference that captures the features relevant to its evaluation, and this formal representation must represent the interrogative act as a finite list of declarative sentences (the possible direct answers to the question being asked). The premisses of a natural-language inference to an interrogative act evoke the interrogative act if the premisses of such a defensible formalization of the inference evoke (in Wi´sniewski’s sense) the formal representation of the interrogative act. Applying this criterion to example 2, we can conclude that the premiss in the headline evokes the interrogative act that follows it.

11.4 Inferences to Open-Ended Questions Wi´sniewski’s account provides a sufficient condition for a valid natural-language inference to an interrogative act, but not a necessary one. Many natural-language inferences from representative illocutionary acts to interrogative acts are intuitively valid but have no defensible formalization that is valid in Wi´sniewski’s sense, because the interrogative act cannot be represented as a request to choose from a finite set of possible direct answers. Almost all why- and how-questions do not meet this requirement; one cannot specify in advance a set of direct answers to such questions. Let us use the term “open-ended question” for natural-language questions that do not contain or entail, even with the help of their context, a specification of their possible direct answers. We cannot apply Wi´sniewski’s formal criteria for evocation to inferences to openended questions. Nevertheless, the informal insights underlying those criteria do seem applicable to them. Despite the absence of a complete finite list of the direct answers to such a question, we can ask whether the premisses in context entail that the interrogative act is sound (i.e. has a true direct answer) and whether they entail a direct answer to it. If the premisses in context do entail that the interrogative act is sound but do not entail a direct answer to it, then the inference from premisses to interrogative act is valid; otherwise, it is invalid. Let us consider how this informally generated conception applies to inferences to open-ended why-questions, open-ended how-questions, and open-ended whatquestions—in each case from a set of representatives articulated in a context.

11.4.1 Why Questions A why-question is a request for an explanation. Such a request is sound, in the sense of having a true direct answer, only if the “thing” for which an explanation is requested is a real phenomenon. The question “Why can’t you put your smart phone down?”, for instance, would not be sound if in fact you can put your smart phone down (i.e.

174

D. Hitchcock

if people did not find their smart phone so addictive). There is no true explanation of a non-existent phenomenon or a non-occurring event. Thus, in a valid argument for a why-question, either the premisses themselves or the context of utterance (or both) must entail the reality of the “thing” for which an explanation is requested. An additional requirement for a sound why-question appears to be that the “thing” for which an explanation is requested actually can be given one. Arguably, some facts cannot be given an explanation; they are “brute facts, random events, things coming about without rhyme or reason, occurrences happening because of nothing in particular, capricious actions, ultimate truths, so’s without why-so’s” (Bromberger, 1992, p. 167). If there are such “things” for which why-questions make no sense, do the premisses and context of a natural-language inference to a why-question need to entail that the “thing” for which an explanation is requested actually has an explanation? Here we can perhaps accept a default position that in general things have explanations, so that a why-question is always apt about any reality, whether this reality is a general phenomenon or a particular event or a particular decision or a person’s current mental state or whatever. An argument for a why-question would thus not need to establish that the “thing” for which an explanation is requested has an explanation, although a critic of such an argument could object that the “thing” in question is the sort of thing that lacks an explanation. The burden of proof would rest on the critic to sustain the objection. In solo reasoning, an inference from a set of beliefs or assumptions to a why-question would be defeasible, subject to a rebutting defeater that makes the question empty.10 The second informal criterion for a valid inference from representative illocutionary acts to an interrogative act is that the premisses do not entail a particular direct answer to the interrogative act. This requirement applies directly to a naturallanguage inference to a why-question: if the premisses in context include or entail a direct answer to the question-conclusion, there is no point in asking it, i.e. the question does not arise. Further, a question does not arise if it is common knowledge, independently of the premisses and context, that there is no point in asking the question. One obvious reason for asking a question is that the questioner does not know the correct answer and is asking the addressee to supply it, as with the question what benefits consciousness brings. Another reason is to motivate interest in finding out the correct answer, in a case where the questioner knows it but the addressee does not, as with the question why you cannot put your smart phone down. There is a similar structure in the pedagogical use of arguments for questions. The teacher can make learners aware of something that needs explanation, as a basis for requesting the explanation, perhaps inviting the learners to make their own suggestions and evaluate the adequacy of those suggestions. In a test situation, an examiner may pose a question to which 10 Bromberger

(1992) on the other hand seems to regard the class of unexplainable facts as quite large, so that judging that a why-question is sound requires knowing that the fact for which an explanation is requested actually has one. Further, as he points out, the structure and content of a why-question do not provide a basis for this knowledge. Hence “what we don’t know when we don’t know why” (the title of his article) is what we must find out to know whether there is anything to know at all (p. 168).

11 Arguing for Questions

175

both the examiner and the test-taker know the correct answer, but the examiner does not know whether the test-taker knows the correct answer and wants to find out. The common feature of all these situations is that posing a question makes sense, whether as a request for information or as a motivator of attention to a subsequent proposed answer or as a stimulus to investigation or as a testing device. An inference to an interrogative act would make no sense if there was no such reason for posing the question. It is debatable, however, whether the existence of a reason for posing a question-conclusion is a requirement for the validity of an inference to it or merely a pragmatic constraint on when drawing the inference makes sense. Let us include it as a requirement for validity, but acknowledge that it may be just a pragmatic constraint. We can summarize the informal requirements for a valid natural-language inference to a why-question as follows: 1. The premisses and context entail that the “thing” for which an explanation is requested is a reality. 2. There is no good reason to think that the “thing” for which an explanation is requested lacks an explanation. 3. The premisses and context do not entail any particular explanation of the “thing” for which an explanation is requested. 4. The context provides a good reason for posing the why-question, typically lack of knowledge of the correct answer by the questioner or the addressees. (This requirement may be a pragmatic constraint on making the inference rather than a requirement for the validity of the inference.)

11.4.2 How Questions How-questions request a description of the mechanism by which something occurs.11 The conditions for a valid inference to a how-question are thus similar to those for a valid inference to a why-question, with the second requirement replaced by the requirement that the “thing” for which a how-explanation is requested is the outcome of, or is constituted by, a process whose components and sequencing can be described. We can summarize them as follows: 1. The premisses and context entail that the “thing” for which a how-explanation is requested is a reality. 2. The “thing” for which a how-explanation is requested is the outcome of, or is constituted by, a process whose components and sequencing can be described. 3. The premisses and context do not entail any particular explanation of the “thing” for which an explanation is requested. 11 Not all interrogative sentences beginning with “how” express a how-question. The sentence “How could you do that?”, for example, is typically not a request for a description of the process that made it possible for the addressee to do “that”. It is a reproach, calling for a justification or an excuse or an apology.

176

D. Hitchcock

4. The context provides a good reason for posing the why-question, typically lack of knowledge of the correct answer by the questioner or the addressees. (This requirement may be a pragmatic constraint on making the inference rather than a requirement for the validity of the inference.)

11.4.3 What Questions Some what-questions are open-ended, such as Wi´sniewski’s example, “What do you know about the case?” (Wi´sniewski, 2013, p. 11), whose utterance in standard contexts would ask the addressee to tell the questioner everything that the addressee knows about the referenced case. How could we apply Wi´sniewski’s two criteria to an inference from representative illocutionary acts to such an open-ended whatquestion? The first criterion is that the premisses entail that the question is sound, i.e. has a correct answer. The soundness of the question “What do you know about the case?” seems to consist solely in there being a single case to which reference is being made. It seems too demanding, however, to require that the set of premisses by itself entails that there is one and only one case, since the identity and reality of the case referred to may be obvious from the context. In general, then, it seems enough to require for the validity of a natural-language inference to an open-ended what-question that the set of premisses and the context of utterance entail the reality and identity of any objects whose reality is presupposed by the interrogative act. Wi´sniewski’s second criterion is that the set of premisses does not entail a particular direct answer to the conclusion. We can embrace this criterion directly. In addition, we should require as before that the context provides a good reason for posing the question. We can sum up the conditions for a valid natural-language inference to an openended what-question as follows: 1. The premisses and the context entail the reality and identity of any objects whose reality is presupposed by the interrogative act. 2. The premisses and context do not entail any particular direct answer to the whatquestion. 3. The context of the inference provides a good reason for posing the what-question, such as the questioner’s or addressees’ lack of knowledge of the correct answer. (This requirement may be a pragmatic constraint on making the inference rather than a requirement for its validity.) It seems possible to extend to other types of open-ended questions the proposed approach to evaluating natural-language inferences to open-ended why-, how- and what-questions. To check the correctness of the proposed criteria for validity of inferences to openended questions, we can apply them to the 11 arguments for open-ended why- and

11 Arguing for Questions

177

how-questions in the 17 examples.12 In six of these arguments, the premiss or premisses entail (or might plausibly be thought to entail) the reality of the phenomenon for which an explanation is requested. Thus they contribute to satisfaction of the criterion of transmission of truth of the premiss to soundness of the question-conclusion. In the other five of these arguments, however, the premisses seem relevant but seem to make no contribution to satisfaction of the criteria derived from Wi´sniewski’s formal account. Each of them is an argument for an open-ended why-question, with a single premiss that neither states nor entails the reality of the phenomenon about which the why-question is asked. A typical example is the previously mentioned argument about smart phones: (4) Your smart phone is making you stupid, antisocial and unhealthy. So why can’t you put it down?

The headline writer makes no attempt to establish that in general people can’t put their smart phone down. Rather, the argument seems to presuppose knowledge by the addressees of the reality of that phenomenon; in other words, the context rather than the premiss entails its reality. Instead of asserting or entailing the reality of the phenomenon for which an explanation is requested, the premiss provides a reason for finding the phenomenon puzzling, and thus in need of an explanation. Smart phones make people stupid, antisocial and unhealthy, so one would expect that people would moderate or stop using them. Hence people’s addiction to them needs explanation. The premiss seems to have the function of establishing the fourth, possibly pragmatic criterion for a good inference to an open-ended why-question: that there is a reason for posing it. The premiss establishes that the phenomenon presupposed by asking the why-question is not what one would expect, and is thus in need of explanation. The smart phone argument is typical in that respect of the five arguments whose premiss does not assert or entail the reality of the phenomenon for which an explanation is requested. Thus, despite initial appearances, the apparent validity of these five arguments is compatible with the proposed account of a valid inference to an openended why-question. The smart phone argument shows that the premisses of a valid inference to a question need not state or entail the presuppositions of the question. However, the context of this argument does include the presupposed information that people “can’t put down” their smart phones—i.e. find it hard to stop using them. One might wonder, therefore, whether any valid inference to a question from one or more representative illocutionary acts must include the presuppositions of the question.13 To answer this question, one needs to clarify the concept of a presupposition of a question. Belnap (1969) defines a presupposition of an interrogative sentence as a declarative sentence

12 There

were also two arguments for what-questions, which however were not open-ended. One was the question “What kind of oil should you cook with?”, which as discussed earlier can be represented as a request to choose from a finite set of direct answers. 13 I thank Marcin Selinger for raising this question in conversation and for further discussion of it in correspondence.

178

D. Hitchcock

that is entailed by each of its direct answers.14 In languages where any declarative sentence entails infinitely many distinct declarative sentences (e.g. those obtained by successively adding one disjunct at a time or by successively conditionalizing with a new antecedent), an interrogative sentence that has one Belnapian presupposition has infinitely many of them. Thus in such languages the premisses and context of an inference to an interrogative sentence cannot assert all its Belnapian presuppositions. They may however entail all of them. Since the falsehood of any such presupposition entails the falsehood of each direct answer (i.e. the unsoundness of the question), the truth of them all is a necessary condition for the soundness of the question. Hence the premisses and context of a valid inference to a question must entail the truth of all its Belnapian presuppositions. But, as the smart phone argument illustrates, they may include other information, such as a reason for thinking that a why-question needs to be investigated—for example, that its obvious direct answer is false. Comprehensive theories of valid inference will need to accommodate valid inferences to interrogative acts. The present author has proposed one such theory, stated as follows: A conclusion follows from given premisses if and only if an acceptable counterfactualsupporting generalization rules out, either definitively or with some modal qualification, simultaneous acceptability of the premisses and non-acceptability of the conclusion, even though it does not rule out acceptability of the premisses and does not require acceptability of the conclusion independently of the premisses. (Hitchcock, 2011, p. 209)

Acceptability in this context is to be construed as an ontic status that in the case of conclusions includes the truth of assertions, the prudence of decisions, the correctness of verdicts, the legitimacy of stipulations, and so on. If the conclusion of an argument is an act of asking a question, then the acceptability of the conclusion is (at least) the soundness of the question. It is doubtful that its acceptability includes ignorance by the addressee of a correct answer to the question, since there are good reasons for asking questions of people who know the answer, such as not knowing the answer oneself and wanting to find it out, or not knowing whether the person knows the answer and wanting to find that out. If one restricts the acceptability of asking a question to the soundness of the question, then one needs only the first of Wi´sniewski’s criteria for a valid inference to a question: transmission of truth into soundness. In Wi´sniewski’s erotetic logic, this criterion is satisfied by purely formal criteria, which hold generally and counter-factually. The proposals in this essay to apply his underlying informal insights directly to natural-language argument to open-ended why-, how- and what-questions also involve appeal to counterfactualsupporting covering generalizations. For example, in general, if an actual state of affairs ϕ has an explanation, then the question “why ϕ?” has a correct direct answer. Thus the position taken in this essay on valid inferences to questions turns out to be a special case of my general conception of good inference. 14 In the logical theory of questions developed by Kubi´ nski (1980), questions of many forms have a

single syntactically constructed presupposition, whose truth entails the question’s soundness. This alternative conception of a question’s presupposition does not apply to open-ended questions, which cannot be represented in the formal language constructed by Kubi´nski.

11 Arguing for Questions

179

11.5 Inferences from Questions to Questions As mentioned at the beginning of Sect. 11.3, Wi´sniewski recognizes a second kind of erotetic inference—namely an inference from an erotetic formula (perhaps supplemented by one or more declarative sentences) to an erotetic formula. The basic insight underlying his account of the validity of such inferences is that one can draw an inference from an initial question to a subsequent question as a way of narrowing down the search space for an answer to the initial question. If one finds a true answer to the subsequent question, then the search for an answer to the initial question becomes easier, because one has to consider fewer direct answers in order to find a true one. We can see how this works with two of Wi´sniewski’s examples. The simpler of the two is an inference from a question to a question: (4) Does John write a lot and does he read a lot? So does John write a lot? (Wi´sniewski, 1996, p. 3) The inferential particle “so” sounds artificial in this example. But one can imagine someone asking themselves the first question and then saying to themselves, “Well, let’s see, does John write a lot?” The initial question has four direct answers, depending on whether each conjunct is answered positively or negatively. The subsequent question has two direct answers: John writes a lot, John does not write a lot. Each of these answers implies that the initial question has two possible direct answers rather than four. That John writes a lot entails that either (1) he writes a lot and reads a lot or (2) he writes a lot but does not read a lot. That John does not write a lot entails that either (3) he does not write a lot but does read a lot or (4) he neither writes a lot nor reads a lot. Thus a direct answer to the question-conclusion “Does John write a lot?” narrows the search space for the initial question. Specifically, either direct answer to the subsequent question leaves only a simpler question to be answered: Does John read a lot? Thus, in this example inferring a question from a question is a way of breaking down into simpler components a search for an answer to a complex question. Wi´sniewski’s second example is an inference from a question and three statements to a question: (5) Is John a monk, or a bachelor, or has he a very patient wife?15 If John is a monk or a bachelor or has a very patient wife, then he is single or married. If John is single, then he is a monk or a bachelor. If John is married, then he has a very patient wife. So is John married or single? (Wi´sniewski, 1996, p. 3) Here the inferential particle “so” sounds (to my ear at least) a bit more natural. Again the inferred question narrows down the search for an answer to the initial 15 This question is the conclusion of a preceding example with two statements as premisses: (1) If John writes three books during one year, then John is a monk or a bachelor or has a very patient wife. (2) John writes three books during one year.

180

D. Hitchcock

question. Given the three statements, then answering the question-conclusion will make it easier to answer the initial question. The answer that John is married implies that the third disjunct of the initial question is correct (and, given definitionally true assumptions, that the first and second disjuncts are false). The answer that John is single implies that either the first or the second disjunct of the initial question is correct (and, given a definitionally true assumption, that the third disjunct is false). Wi´sniewski proposes the following two conditions as individually necessary and jointly sufficient for the validity of an inference from an erotetic formula and a (possibly empty) set of declarative sentences to an erotetic formula: (Transmission of soundness/truth into soundness): If the initial question is sound and all the declarative premisses are true, then the question which is the conclusion must be sound. (Open-minded cognitive usefulness): For each direct answer B to the question which is the conclusion, there exists a non-empty proper subset Y of the set of direct answers to the initial question such that the following condition holds: (♣) if B is true and all the declarative premisses are true, then at least one direct answer A ∈ Y must be true. (Wi´sniewski, 2013, pp. 52–53)

These statements of the conditions are informal, to be sharpened later with reference to concepts defined for erotetic formal languages. Let us call the two conditions, for short, transmission and usefulness. In an inference that satisfies the transmission condition, the truth of the declarative premisses (if any) and the soundness of the question-premiss entail the soundness of the question-conclusion. Soundness of a question, as before, means that it has at least one true direct answer. An inference that satisfies the usefulness condition is cognitively useful in the sense that acquiring knowledge that a certain direct answer to the question-conclusion is true helps to acquire knowledge of which direct answers to the question-premiss are true, because the truth of a direct answer to the question-conclusion (along with the truth of any declarative premisses) implies that there is a smaller set of direct answers to the initial question within which an answer is true. This cognitive usefulness is open-minded in the sense that it does not pronounce on which direct answers to the questionpremiss are true, since it does not pronounce on which direct answers to the questionconclusion are true. As with erotetic inferences of the first kind, Wi´sniewski’s formal account of the validity of erotetic inferences of the second kind can be applied only to naturallanguage inferences from and to questions that can be represented as a request to select from a finite set of possible direct answers. Thus his formal apparatus can be used only for some natural-language inferences from questions to questions. However, just as with erotetic inferences of the first kind, one can apply the informal insights underlying his formal account directly to natural-language inferences from and to open-ended questions that cannot be represented as a request to select from a finite set of possible direct answers. In such direct applications to natural-language inferences, one must allow as before for consideration of the context in determining whether the two conditions are met. The original 17 examples included only one natural-language inference from a question to a question:

11 Arguing for Questions

181

(6) What value does conciousness [sic] add, and therefore why did it ever come into existence?16

A search of the Web using “so why”, “how then” and “therefore why” as search terms in combination with a content word (machine, fruit, broccoli, lights, emotional, happiness) turned up many examples of natural-language inferences to questions from statements, but not a single example of a natural-language inference from a question to a question. If Web pages have the same range of structures as human spoken discourse and written texts in general, then humans rarely articulate inferences from questions to questions. Let us however consider the one example that has turned up. The author of example (6) is clearly asking the first question: what value does consciousness add? That is the question that respondents on Quora answered, and the author raised no objection to their interpretation. The author seems to assume that the correct answer to the first question will help to answer the second question. The assumption that knowing the value that consciousness adds will tell us why it came into existence belongs to an adaptationist paradigm according to which all genetically caused characteristics which belong to most members of a biological species have been acquired as a way of helping them to adapt to their environment, in the sense of increasing the chance that there will be fertile offspring with the genetic basis of the characteristic. The reasoning is flawed, in that characteristics of organisms can come into being as accidental consequences of something else that is selected for in the process of evolution (or just as accidental changes), and these accidentally acquired characteristics may be helpful, harmful or neutral (Gould & Lewontin, 1979). Thus the fact that a genetically caused characteristic is beneficial to an organism does not show that it came into being as a result of selection for that benefit in the process of evolution. At best, it establishes that the characteristic may have come into being as a result of evolutionary pressure, i.e. that the hypothesis of such an origin is worth investigating. In any case, the author of example (6) is not inferring a question from a question as a way of narrowing down the search for an answer to the initial question. Rather, the author is asking the first question as a way of narrowing down the search for an answer to the second question. Thus the supposed relationship of the two questions is the reverse of the relationship in a valid erotetic inference of the second kind. The word “therefore” indicates a future inference from an answer to the first question to an answer to the second question, rather than a present inference from a broader question to a narrower question. Wi´sniewski’s account of the validity of erotetic inferences of the second kind thus appears to have little applicability to articulated human inferences. He reports, however, that it has turned out to be applicable to problem solving and to proof theory (Wi´sniewski, 2013, p. 2). Thus, even though humans do not articulate inferences from a question to a question, the account of the validity of such inferences is helpful as 16 The

question was posted in September 2012 on Quora, a website on which anyone can submit a question and invite answers from volunteers. It is available at https://www.quora.com/What-valuedoes-conciousness-add-and-therefore-why-did-it-ever-come-into-existence; accessed 2019 01 10.

182

D. Hitchcock

an aid to research. It gives guidance on deriving from an initial complex question simpler questions whose true answers narrow down the search.

11.6 Summary People sometimes argue for questions, as one can easily establish by a Web search using strings such as “so why” and “how then” that consist of a conclusion indicator and an interrogative particle, in either order. The texts thus discovered are arguments, in the sense that they are most plausibly interpreted as providing a reason for asking the question-conclusion. And the conclusion is a genuine request for information, not a merely rhetorical question. Such arguments have as their immediate goal to put the question “on the table” as something that deserves to be answered. Putting it on the table can be a motivation for a reader to read on to find out the answer that the writer provides. Or it can be a challenge to addressees or to a third party to justify or excuse some behaviour. In educational contexts, it can have pedagogical or examination purposes. The “inferential erotetic logic” developed by the Polish logician Andrzej Wi´sniewski provides a basis for evaluating the inferences in such arguments. This logic applies to formal languages with two types of sentences: declarative sentences that have a truth-value when interpreted and “erotetic formulas” that express a question as a finite set of two or more declarative sentences and that are sound when interpreted if and only if at least one declarative sentence in the set is true. Wi´sniewski’s logic is directly applicable to natural-language inferences to questions only if the question-conclusion can be represented as a request to select from a finite set of possible direct answers. However, the informal insights underlying his logic can be applied to natural-language inferences to open-ended questions that cannot be represented as a request to select from a finite set of possible direct answers. Wi´sniewski distinguishes two kinds of inferences to an erotetic formula. According to Wi´sniewski, an inference to an erotetic formula from one or more declarative sentences is valid if and only if the set of declarative sentences entails the soundness of the erotetic formula without entailing any particular direct answer. These two conditions can be applied directly to natural-language inferences from a set of representative illocutionary acts to an open-ended question. Such inferences are valid only if the premisses and context of the inference entail the reality of what is asked about and the legitimacy of what is asked about it without entailing a particular answer to the question. In addition, perhaps as a pragmatic constraint rather than a condition for validity, there must be a point to asking the question (such as the addressees’ ignorance of its true direct answers). The premisses of such inferences typically either assert the reality of what is asked about or establish that there is a point to asking the question. According to Wi´sniewski, an inference from an erotetic formula and a possibly empty set of declarative sentences to an erotetic formula is valid if and only if the soundness of the initial question and the truth of the declarative sentences jointly

11 Arguing for Questions

183

entail the soundness of the question-conclusion and the question-conclusion narrows down the search for an answer to the initial question (in the sense that each answer to the question-conclusion entails along with the declarative premisses that a non-empty proper subset of its direct answers contains a true answer to the initial question). A Web search turned up no natural-language inferences from a question to a question, an indication that humans generally do not articulate such inferences. Their logic is however applicable to problem solving and proof theory, where systematic decomposition of an initial question can make it easier to answer it.

References Belnap, N. D., Jr. (1969). Questions, their presuppositions and how they can fail to arise. In K. Lambert (Ed.), The logical way of doing things (pp. 23–37). New Haven: Yale University Press. Bromberger, S. (1992). On what we don’t know when we don’t know why. On what we know we don’t know: Explanation, theory, linguistics, and how questions shape them (pp. 145–169). Chicago and London: University of Chicago Press. Fennessy, J., Bidon, T., Reuss, F., Kumar, V., Elkan, P., Nilsson, M. A., et al. (2016). Multi-locus analyses reveal four giraffe species instead of one. Curr. Biol., 26, 2543–2549. Gould, S. J., & Lewontin, R. C. (1979). The spandrels of San Marco and the Panglossian paradigm: A critique of the adaptationist programme. Proceedings of the Royal Society of London, B, 205, 581–598. Groenendijk, J., & Stokhof, M. (1996). Questions. In J. van Benthem & A. ter Meulen (Eds.), Handbook of logic and language (pp. 1055–1124). Amsterdam: Elsevier. Hitchcock, D. (2011). Inference claims. Informal Logic, 31(3), 191–229. Hitchcock, D. (2018). The concept of an argument. In S. Oswald & D. Maillat (Eds.), Proceedings of the European Conference on Argumentation 2017, (Vol. II, pp. 361–374). London: College Publications. Hitchcock, D. (2019). We justify questions, so how does that work? (Forthcoming) In B. Garssen, J. Wagemans, G. Mitchell, & D. Godden (Eds.), Proceedings of the Ninth International Conference on Argumentation (pp. 28–43). Amsterdam: SICSAT. Available at http://cf.hum.uva.nl/issa/ ISSA_2018_proceedings.pdf; accessed 2019 08 27. Kraus, M. (2006). Arguing by question: A Toulminian reading of Cicero’s account of the enthymeme. In D. Hitchcock & B. Verheij (Eds.), Arguing on the Toulmin model: New essays in argument analysis and evaluation (pp. 313–325). Dordrecht: Springer. Kubi´nski, T. (1980). An outline of the logical theory of questions. Berlin: Akademie-Verlag. Lumer, C. (1990). Praktische Argumentationstheorie: Theoretische Grundlagen, praktische Begründung und Regeln wichtiger Argumentationsarten (Practical theory of argumentation: Theoretical foundations, practical justification and rules for important types of argumentation). Braunschweig/Wiesbaden: Vieweg. Lumer, C. (1991). Structure and function of argumentations—An epistemological approach to determining criteria for the validity and adequacy of argumentation. In F. H. van Eemeren, R. Grootendorst, J. A. Blair, & C. A. Willard (Eds.), Proceedings of the Second International Conference on Argumentation (pp. 98–107). Amsterdam: SICSAT. Searle, J. R. (1976). A classification of illocutionary acts. Language in Society, 5(1), 1–23. Van Eemeren, F. H. (1987). For reason’s sake: Maximal argumentative analysis of discourse. In F. H. van Eemeren, R. Grootendorst, J. A. Blair, & C. A. Willard (Eds.), Argumentation: Across the Lines of Discipline: Proceedings of the Conference on Argumentation 1986 (pp. 201–215). Dordrecht, Holland: Foris.

184

D. Hitchcock

Van Eemeren, F. H., & Grootendorst, R. (1984). Speech acts in argumentative discussions: A theoretical model for the analysis of discussions directed towards solving conflicts of opinion. Dordrecht, Holland: Foris. Van Eemeren, F. H., & Grootendorst, R. (1992). Argumentation, communication, and fallacies: A pragma-dialectical perspective. Hillsdale, NJ: Lawrence Erlbaum Associates. Van Eemeren, F. H., & Grootendorst, R. (2004). A systematic theory of argumentation: The pragmadialectical approach. Cambridge: Cambridge University Press. Wi´sniewski, A. (1996). The logic of questions as a theory of erotetic arguments. Synthese, 109(1), 1–25. Wi´sniewski, A. (2013). Questions, inferences and scenarios. London: College Publications.

Chapter 12

Expressives in Argumentation: The Case of Apprehensive Straks (‘Shortly’) in Dutch Ronny Boogaart

12.1 Introduction Dutch straks (‘shortly’) is a temporal adverb used to refer to the near future. As such it functions in most instances as a predicate modifier, situating an event in time relative to the moment of speaking. However, the word may also be used as part of an expressive pattern presenting an event that may possibly happen but that the speaker evaluates negatively (Boogaart, 2009), as in (1).1 (1) Pas op! Straks val je! ‘Watch out. You might fall.’2 The second sentence here is not just describing an event that may occur in the near future, but rather presenting an event that the hearer is supposed to prevent from happening at all. The semantics of the construction combines epistemic modal meaning, i.e. possibility, and a sense of undesirability and it may therefore be identified as an “apprehensional epistemic” in the sense of Lichtenberk (1995), typically used for the illocutionary act of warning, as in (1) (Angelo & Schultze-Berndt, 2016). In this chapter, a linguistic, construction-based analysis of apprehensive straks is combined with an account of its use in argumentative discourse as a presentational device in strategic maneuvering (van Eemeren, 2010). Speakers typically use the construction to provide an argument in favor of doing something by pointing out R. Boogaart (B) Leiden University Centre for Linguistics, Leiden, The Netherlands e-mail: [email protected] 1I

use the term expressive for utterances that express, rather than describe, the speaker’s emotion (see Sects. 12.2.1 and 12.4.1) (cf. Potts, 2007; Foolen, 2015). This class of utterances does not coincide with Searle’s (1976) category of expressive speech acts. 2 More specific argumentative uses of the Dutch construction to be discussed in Sect. 12.3 may be more literally rendered in English by pretty soon, next or next thing (we/you know), as in: let me guess, next you will say the moon landing was a hoax? In many other instances, however, there is no real equivalent for Dutch straks in the English translation of the examples. © Springer Nature Switzerland AG 2020 F. H. van Eemeren and B. Garssen (eds.), From Argument Schemes to Argumentative Relations in the Wild, Argumentation Library 35, https://doi.org/10.1007/978-3-030-28367-4_12

185

186

R. Boogaart

the negative consequences not doing it would have. This is already apparent in the warning use exemplified in (1): the hearer is urged to watch out because if (s)he does not watch out (s)he might fall. Use of the Dutch construction is, however, not restricted to such contexts of immediate danger and it extends well beyond the contexts that are given in the typological literature to illustrate the category of apprehensive (see especially Angelo & Schultze-Berndt, 2016). A type of use that is probably furthest removed from the concrete warning use in (1) is illustrated in (2), where the temporal meaning of straks has bleached since the construction does not present a future event. The speaker uses the construction to present an absurd standpoint that he himself obviously does not agree with, as is clear from the surrounding context. In this post from 2008, the speaker comments on a remark by the lawyer of Josef Fritzl that his client should not go to jail because he may be mentally ill. (Fritzl, from Austria, was arrested for having kept his daughter hostage for 24 years and making her pregnant seven times in this period.) (2) Te gek voor woorden. Straks is hij nog zielig en het slachtoffer en de dochter de dader. Ja, ja, sta nergens meer van te kijken. www.nujij.nl/advocaat-wil-fritzl-niet-in-gevangenis.2501475.lynkx ‘This is too crazy for words. Next [they will say that] we have to feel sorry for him and he is the victim and the daughter is the perpetrator. Yes, really, nothing surprises me these days’ Such argumentative uses of apprehensive straks share their expressive nature and negative evaluation with the more everyday warning use in (1), but they crucially differ from it in other respects. Outside of contexts of concrete warning and advice, apprehensive straks is used almost exclusively in fallacious argumentative moves, incorrectly applying causal or analogical argumentation, as in (2). From the viewpoint of strategic maneuvering, this raises an interesting question: If arguers try to balance between reasonableness and effectiveness, then why would they use such an expressive stylistic device that is typically and recognizably presenting a derailed maneuver? More generally, the analysis of an expressive construction like apprehensive straks not only adds to the growing body of literature providing a linguistically informed account of strategic stylistic choices in argumentation (e.g. Snoeck Henkemans, 2009, 2013; Fahnestock, 2011; Boogaart, 2013; van Leeuwen, 2014; Jansen, 2017; van Haaften & van Leeuwen, 2018), but it may also contribute to ongoing debates about the role of emotion in linguistics (Foolen, 2015; Corver, 2016) and in argumentation theory (summed up by Krabbe & van Laar, 2015, pp. 1178–1186). In this chapter, I will first present formal and semantic properties of the linguistic pattern in (1) and (2), in order to substantiate the claim that it should be considered an independent construction expressing an emotion (Sect. 12.2). In Sect. 12.3, I will distinguish between three different usages of the construction in argumentative discourse. In Sect. 12.4, it will be shown that, for several reasons, presenting an argument by means of an expressive may be considered an “immunization strategy” (van Eemeren & Grootendorst, 1992, p. 119).

12 Expressives in Argumentation: The Case of Apprehensive …

187

12.2 Form and Meaning of the Apprehensive Straks-Construction If an utterance containing a temporal adverb referring to the future such as straks is interpreted as an argument for or against a standpoint, this does not in itself warrant the conclusion that we are dealing with a construction, i.e. a pairing of a specific form and a specific meaning. The argumentative status of such an utterance could be determined fully compositionally on the basis of the temporal semantics of straks and the specific argumentative context it is used in. However, the pattern with straks that was exemplified in (1) and (2) has some idiosyncratic formal and semantic properties that make it different from other clauses containing the temporal adverb straks. These properties suggest that the pattern with straks is an independently learnt and stored unit, i.e. a construction in the sense of construction grammar (e.g. Tomasello, 2003; Hilpert, 2014), with an intrinsic “argumentative” meaning in the sense of Anscombre and Ducrot (1983), Ducrot (2009) and Verhagen (2005). In this section, I will show that the pattern indeed constitutes an expressive construction in its own right (12.2.1) and then address its argumentative meaning (12.2.2).

12.2.1 Apprehensive Straks as a Construction The difference between the “normal”, temporal use of straks and the apprehensive straks-construction can be clearly demonstrated by a pair such as (3)–(4). (3) De winkel is straks gesloten. ‘The shop will be closed shortly.’ (4) Straks is de winkel gesloten! ‘The shop might be closed!’ In (3), the speaker asserts that the shop will be closing in a little while. Given an argumentative context, the utterance may be interpreted as an argument, for instance for hurrying up, but there is nothing inherently evaluative or argumentative about the phrasing of (3). By using the apprehensive straks-construction in (4), on the other hand, the speaker expresses his fear that the shop might be closed. Thus, the utterance has an expressive dimension (Potts, 2007) that is lacking in (3). This is highlighted by a specific sentence melody, with a higher pitch, a rising intonation contour and a faster speech rate.3 In fact, if (4) is pronounced with a neutral intonation contour— or, in written language, if the exclamation mark is lacking—then it may simply be interpreted in the same way as (3). The only difference would then be that in (4) the temporal adverb is in clause-initial position. In that case, both (3) and (4) would be simple assertions. On the apprehensive use of (4), however, the utterance is not 3 Prosodic

properties of utterances are treated here as a presentational device even in written discourse, in accordance with Chafe’s (1988) concept of “covert prosody”.

188

R. Boogaart

a neutral description but the expression of an emotion and this is supported by its “expressive” prosodic properties. What exactly the expressive component of the construction amounts to semantically is hard to make precise, but this is a actually a typical property of expressives, named “descriptive ineffability” by Potts (2007, p. 166): “Speakers are never fully satisfied when they paraphrase expressive content using descriptive, i.e. nonexpressive terms”. Thus, (4) expresses something else, or something more, than can be captured by describing this emotion, as in ‘I am afraid that the shop might be closed’. In this chapter, I will keep on using the term apprehension to describe the kind of emotion that is expressed by speakers who use the construction. In fact, Angelo and Schultze-Berndt (2016, p. 283) explicitly include the Dutch construction in this cross-linguistic category. Apart from the additional expressive dimension, even the descriptive content of (4) might be different from that in (3). Crucially, in (4), the speaker’s apprehension may actually be about the shop being closed at the moment of speaking, i.e. the utterance does not need to concern the future. For instance, the utterance may be used on a Sunday to express the possibility that the shop is closed on Sundays, thus including the moment of speaking itself. The same thing holds for the example that was given in (2): the idea that ‘the perpetrator is really the victim’ is not a proposition that may become true in the future, but rather an extreme idea that is ascribed to an antagonist at the moment of speaking (see Sect. 12.3.2). This clearly shows that, in the apprehensive construction, straks is not a regular temporal adverb situating a state or event in the immediate future of the time of speaking like it does in (3). In fact, in using the apprehensive construction, the speaker may even express his premonition that an event happened in the past, as in (5), where the adverb is combined with a present perfect to present a possible event preceding the point of speech. (5) ik denk straks is i.e. in de sloot gereden (Corpus of Spoken Dutch fn007974.148) ‘I’m like, he might have driven into the ditch!’ Whereas warnings, as in (1), are by definition about future events that should be prevented from happening, the Dutch construction, more generally, expresses apprehension that something undesirable might be, or might become, true (in past, present or future). This does not mean that in this construction the temporal sense of straks has disappeared altogether: what is situated in the future in (4) and (5) is not the time of the event itself, but rather some point of evaluation at which it may turn out that the state is holding or that the event did in fact happen in the past (Boogaart, 2009, pp. 170–174). At the moment of speech itself, the speaker cannot yet be sure about this, which constitutes the epistemic-modal component of the semantics of the construction, i.e. possibility.4 Use of the construction in (2) is slightly different but 4 This is highly comparable to the way epistemic modal meaning may develop in auxiliaries that are

used to indicate future time. Thus, Sweetser (1990, p. 62) paraphrases That will be John (uttered upon hearing the hone ringing) as “my present theory that that is John will proceed to future verification/confirmation”.

12 Expressives in Argumentation: The Case of Apprehensive …

189

here as well we may assume a point of evaluation in the future, constituted by the time at which the antagonist may voice the absurd idea that the perpetrator is really the victim and vice versa. This speech event in the near future may be made explicit, and in fact it often is (cf. English Next (thing) you/they will say that…), but it may also remain implicit as is the case in (3).5 That apprehensive straks is not modifying the time of the event in the way a regular temporal adverb would can also be seen in sentences such as (6) that contain a temporal adjunct in addition to apprehensive straks. (6) Straks zitten we hier volgend jaar nog! (www.context.reverso.net) ‘We could still be here next year’ Moreover, by using volgend jaar (‘next year’) and nog (‘still’) the speaker is stressing that the situation may last for a long time and this is incompatible with the temporal sense of straks (“immediate future”). Since temporal straks and apprehensive straks are used to express different things, they might, in principle, even occur in one and the same sentence. An example of this is given in (7). (7) Straks ga je straks je volgende vissen halen (…) 100 km verder, kom je thuis… binnen 3 dagen je nieuwe vissen dood door de Julido’s. (https://aquaforum.nl/threads/200l-hoe-ga-ik-die-vullen.42571) ‘Next thing you will be getting new fish at a 100 km distance, then you arrive home with them and within three days your new fish are killed by the Julido’s’ The writer of this post argues against keeping Julido fish in a home aquarium by sketching a horror scenario that might happen if new fish are introduced in the aquarium. In the first sentence, he uses the adverb straks twice: the first one is an instance of apprehensive straks, marking the whole scenario as something to be feared; the second one seems to be an ordinary predicate modifier, like volgend jaar (‘next year’) in (6), situating the first event in the near future. The examples in (6) and (7) show that the scope of apprehensive straks is larger than just the predicate, marking the whole utterance as an expressive. In (6), it has scope over the predicate modifier volgend jaar and in (7) it even has scope over more than one clause. The event to be feared and to be prevented in (7) is the one in the final clause (i.e. the fish being killed) more than the rather innocent activity in the first clause (i.e. going out to get new fish). In fact, using apprehensive straks only for the first step in the scenario does not make sense (see (8)), which is further evidence for the claim that the scope of apprehensive straks in (7) includes more than one clause.6 (8) ??Straks ga je straks je volgende vissen halen! ‘Next thing you will be getting new fish!’ 5 It is my impression that it is much harder to leave out explicit reference to this future speech event

with English next (thing) than it is with apprehensive straks in Dutch. is quite easy to find more instances of apprehensive straks covering such a scenario, which in fact seems to constitute a sub-construction of its own, of the form ‘straks [x and then y]!’. In such examples, the possible occurrence of y is often used an argument against doing x. In (7), however, the scenario as a whole is used as an argument against getting the Julido fish. 6 It

190

R. Boogaart

So why treat this pattern as a construction in the sense of construction grammar rather than say that the word straks has a temporal and a modal sense, each with their own syntactic scopal properties, or that the apprehension reading of straks follows from the context? For one thing, as noted above, the apprehension reading only arises if the utterance has specific prosodic properties and, in addition to straks itself, these should thus be considered part of the form of the construction. Indeed, in (7), the rising intonation contour extends all the way up to the end of the third clause. Moreover, the construction is more general than the adverb straks: the Dutch apprehension construction requires an adverb referring to immediate future in first position.7 Straks is the most frequent and the most standard one, but in fact many adverbs that in different varieties of Dutch are used to indicate immediate future can be used in the construction.8 It makes more sense, therefore, to treat the apprehension reading as a property of the pattern, with an adverb indicating “immediate future” in first position, rather than assume that, in Dutch, all words indicating immediate future happen to be polysemous or ambiguous between the temporal and the apprehension reading. Besides, the semantic contribution of the temporal adverb straks to the apprehension construction may be bleached, since it not necessarily modifies the time of the event itself, but even in such cases, as in (2), (4) and (5), it has a temporal sense to it and it does not by itself express apprehension. The observation that different temporal adverbs expressing immediate future may occupy the first position in the construction in fact suggests that this specific temporal meaning still does contribute to the meaning of the construction as a whole. Presumably, just like the prosodic properties, the sense of “immediacy” in adverbs such as straks, dadelijk and direct (all translatable as ‘shortly’ or ‘pretty soon’) adds to a feeling of “urgency” and thus to the expressive nature of the construction. To sum up, assuming that constructions are pairings of a specific form and a specific meaning, we may characterize the apprehensive straks construction, on the form side, as a pattern with an adverb indicating immediate future in first position as well as expressive prosodic properties (high pitch, rising intonation, fast speech rate). On the semantic side, straks (nog) p! expresses the speaker’s fear that p might be or might become true. As a result, the construction is typically used in an argumentative way to point out negative consequences and it may be considered an instance of “grammaticalized” argumentativity in the sense of Ducrot’s (2009) “argumentative semantics”.

7 The

first position of the adverb is certainly typical for the construction, as is the use of scalar particles such as nog (‘still’), but there are exceptions (discussed by Boogaart, 2009). 8 Including Zo meteen, (zo) direct, da(de)lijk, seffens, v(o)ort, aans, d’rmee/temee (all translatable as ‘shortly’).

12 Expressives in Argumentation: The Case of Apprehensive …

191

12.2.2 Argumentative Semantics Since speakers use apprehensive straks to express a negative attitude towards a situation, it is tempting to say, as I did earlier (Boogaart, 2009), that the construction has an intrinsic “negative argumentative orientation” in the sense of Ancombre and Ducrot (1983), Ducrot (2009) and Verhagen (2005). However, the concept of “argumentative orientation” is not really used by these authors for matters of attitude or evaluation, but rather for matters of polarity (affirmative/negative). Therefore, it is more accurate to say that apprehensive straks has a negative evaluative, rather than argumentative, orientation. The difference between evaluative and argumentative orientation can be illustrated by means of almost and barely, that have a different argumentative orientation (Verhagen, 2005, pp. 45–55). Thus, the use of barely (passing) in (9) triggers conclusions that follow from not (passing), whereas the use of almost in (10) rather orients the reader towards conclusions that follow from passing the exam. (9) Our son barely passed the exam. (10) Our son almost passed the exam. In these examples, the argumentative orientation of barely (negative) and almost (positive) “overrules” the descriptive content. According to (9), our son did actually pass the exam, whereas according to (10) he did not, but by using the “argumentative operators” barely and almost, the speaker directs the hearer’s attention to the possibility of not passing in (9) and to the possibility of passing in (10). Now, in (9) and (10) positive and negative argumentative orientation (in the yes/no sense) corresponds to a positive and a negative evaluation, since passing an exam is considered something to be applauded, but this is by no means necessary, as (11) and (12) make clear. (11) Our son barely failed the exam. (12) Our son almost failed the exam. Barely and almost have an intrinsic, context-independent negative versus positive argumentative orientation, but in terms of evaluation, this works out the opposite way in (11) and (12) as compared to (9) and (10). Applied to apprehensive straks this means that it does not have a negative argumentative orientation, but rather a negative evaluative orientation. The latter term is used by Verhagen (2000) for the modal use of the verb dreigen (‘to threaten’) in sentences such as (13). (13) Het dreigt slecht weer te worden. ‘The weather threatens to turn bad’ (14) Het belooft mooi weer te worden. ‘The weather promises to be good’ When used as a modal auxiliary, as in (13) and (14), both dreigen (‘to threaten’) and beloven (‘promise’) present the speaker’s expectation about the future; the difference

192

R. Boogaart

between the two is that the latter one triggers positive conclusions, whereas the first one rather tells the hearer to take necessary precautionary measures. The difference between the two is one of evaluation, rather than polarity, and in this sense apprehensive straks clearly shares the negative evaluative orientation of dreigen.9 In fact, the inferences Verhagen (2000, p. 202) assumes to be attached to dreigen (“Be prepared”, “Do something”) are often applicable to apprehensive straks as well. As for argumentative orientation in the polarity sense, apprehensive straks, like both dreigen and beloven, is positive since it presents the possibility that a situation might happen or might be true and it invites the hearer to draw inferences from that possibility. Since the apprehensive straks construction systematically constrains the kind of inferences that the hearer is supposed to draw from an utterance, it functions as a kind of “grammaticalized” argumentativity, as Ducrot and Verhagen assume for various connectors, adverbs and grammatical constructions (see Boogaart & Reuneker, 2017 for an overview). In the remainder of this chapter, I will show how this property works out if the construction is used in argumentative discourse. More specifically, the apprehensive straks-construction will be described as a presentational device for strategic maneuvering (van Eemeren, 2010). In Sect. 12.3, I will make a distinction between three different uses of the construction in argumentative discourse; in Sect. 12.4, I will address the question in what way the use of an expressive construction like apprehensive straks is a strategic presentational device.

12.3 Apprehensive Straks in Argumentative Discourse In the preceding section, the semantics of apprehensive straks was described as a combination of epistemic modality and an expressive dimension of apprehension: the speaker fears that p might be, or might become, true. In terms of Verhagen (2000) we may rephrase this as a combination of a positive argumentative orientation (“possibility”) and a negative evaluative orientation. In this section, I will show that these semantic properties make the construction well suited to be used in the following three argumentative contexts: (i) warnings; (ii) slippery slope argumentation10 ; (iii) Reductio ad Absurdum (RAA).

9 This

raises the question if there is also a positive counterpart for straks (‘shortly’) like there is for dreigen (‘to threaten’). An interesting candidate is sentence-initial wie weet (‘who knows’) (Boogaart, 2017); there is empirical evidence for the claim that it is systematically used to present possible situations that are evaluated positively (van Ginkel, 2018). 10 The term is used here in a neutral sense, i.e. not necessarily as referring to a fallacy (Walton, 1992, p. 229), but it will turn out that when presented by means of apprehensive straks it is usually overtly fallacious (see especially Sect. 12.4.2).

12 Expressives in Argumentation: The Case of Apprehensive …

193

Since both (i) and (ii) involve causal argumentation they will be treated together in Sect. 12.3.1; in Sect. 12.3.2, I will analyze the use of apprehensive straks in (iii) as an argument from comparison.

12.3.1 Pragmatic Argumentation: Warnings and Slippery Slope Adverbs referring to the future may, in general, be indicative of causal argumentation since this type of argumentation often concerns future events (van Eemeren, Houtlosser, & Snoeck Henkemans, 2005, p. 214): a course of action in the present may be advised, or advised against, by pointing out positive or negative consequences in the future. Since apprehensive straks, more specifically, presents a negative attitude towards a situation, the construction is naturally used in arguments about undesirable consequences, i.e. in a specific kind of pragmatic argumentation (van Eemeren & Snoeck Henkemans, 2017, p. 89). As such, apprehensive straks may be used to present either the argument in (15a) or the one in (15b).11 (15)

a

b

X is undesirable

X is desirable

Because: X leads to Y

Because: not-X leads to Y

And Y is undesirable

And Y is undesirable

Thus, if the apprehensive straks-construction is used as a presentational device in pragmatic argumentation, it can only be used to present negative consequences (Y), but the argument may be either an argument against X (in 15a) or in favor of X (in 15b). Use of the construction in warnings could go either way: the hearer is urged to do something or to stop doing something in order to prevent Y from happening. Whereas in (16) the hearer is explicitly told to watch out (X is desirable) because if (s)he does not watch out (s)he might fall (=15b), in (17) (s)he is told to stop caressing the animal (X is undesirable) because, if (s)he continues to do so, it might bite (=15a). (16) Pas op! Straks val je! ‘Watch out. You might fall.’ (17) Aai dat beest nou niet, zegt mama, straks bijt i.e. 11 In this respect, it may be compared to the use of anders (‘otherwise’) (cf. van Eemeren, Houtlosser, & Snoeck Henkemans, 2005, p. 214). However, anders (‘otherwise’, ‘if not’) lacks the expressive dimension of apprehensive straks.

194

R. Boogaart

(Corpus of Spoken Dutch fn001132.176) ‘Now don’t stroke that animal, says mum, it might bite.’ By using straks p following a prescriptive (or ‘practical’) standpoint presented in the form of a directive speech act,12 the speaker is clearly providing an argument in favor of this standpoint and such a move is closely following the general scheme for pragmatic argumentation on the basis of negative consequences given in (15). The second type of argumentative use for straks, i.e. slippery slope argumentation, equally instantiates the argumentative scheme in (15), except here Y is not an immediate consequence of X itself, but rather the climax in a chain of negative consequences triggered by X. A typical example is given in (18), a line from the song Het hondje van Dirkie (‘Dirkie’s little dog’), written by Louis Davids in 1925. (18) Van wie is dat stuk gespuis? Straks heb ik Artis in me huis! ‘Who does that piece of scum belong to? Pretty soon I will have Artis [the Amsterdam Zoo] in my house!’ The lyrics tell the story of the young boy Dirkie who one day witnesses a car knocking down a dog. He takes the dog home and takes care of it there, but this should be kept a secret from his mother who does not want animals in her house. Then one day the dog hobbles into the living room and Dirkie’s mother yells out (18). Her argument for the standpoint that it is undesirable for the dog to be in the house, indirectly communicated by the preceding question (‘who does that piece of scum belong to?’), is that allowing one animal in her house, in the long run, will lead to the entire Amsterdam Zoo living there. This is clearly an instance of fallacious slippery slope argumentation, a common abuse of the pragmatic argumentation scheme in (15), wrongly suggesting “that adopting a certain course of action will inevitably go from bad to worse when in fact there is no evidence that such an effect will occur” (van Eemeren & Snoeck Henkemans, 2017, p. 117). Whereas slippery slope argumentation as such is not necessarily fallacious, it seems that the apprehensive straks-construction lends itself very well as a stylistic device for presenting this particular fallacy. It is easy to find examples in current discussions about, for instance, the reception of refugees (see (19)) or extending laws on euthanasia (see (20)). (19) Straks worden wij nog asielzoekers. Alleen jammer dat we nergens terecht kunnen. (twitter @EvertGDavelaar 22/6/2017) ‘Next thing we will become asylum seekers ourselves. Too bad there will be no shelter for us.’ (20) Straks komen we nog in een wereld waarin men zegt: O, jij bent al over de 70, wordt het niet eens tijd dat je verhuist naar de andere wereld? (https://www.startpagina.nl/v/gezondheid/vraag/51464/moeilijk-waardigemanier-sterven-wilt) ‘Pretty soon we will arrive into a world in which they say: since you are over 70, is it not about time for you to move to the other world?’ 12 The

standpoint need not be explicit and then the apprehensive straks-constructions functions as a warning by itself. This, in fact, is typically true of apprehensives as a cross-linguistic category (Angelo & Schultze-Berndt, 2016).

12 Expressives in Argumentation: The Case of Apprehensive …

195

Once a linguistic device is so intimately tied up with a particular fallacy, one may wonder how it can still be an effective choice when strategically maneuvering between reasonableness and effectiveness, since strategic maneuvering is all about being effective without noticeably giving up on standards of reasonableness. I will address this issue in Sect. 12.4.

12.3.2 Reductio ad Absurdum A third type of use that is typical for apprehensive straks as a presentational device in argumentation less clearly constitutes an instance of the schema in (15) since there is no temporal sequence and causality involved, at least not literally. However, the type does seem to be a natural extension of fallacious slippery slope argumentation of the kind illustrated in (18), (19), and (20). An example of the third type was given in (2), repeated here as (21). (21) Te gek voor woorden. Straks is hij nog zielig en het slachtoffer en de dochter de dader. Ja, ja, sta nergens meer van te kijken. (www.nujij.nl/advocaat-wil-fritzl-niet-in-gevangenis.2501475.lynkx) ‘This is too crazy for words. Next [they will say that] we have to feel sorry for him and he is the victim and the daughter is the pepetrator. Yes, really, nothing surprises me these days’ Another example is the headline in (22), presenting a quote from the vice-president of FC Barcelona, Vilarrubi, who is responding to allegations about fraudulent acts surrounding the transfer of soccer superstar Neymar from Santos to FC Barcelona. (22) Preses Barcelona: “Straks is Neymar nog betrokken bij dood JFK” (https://www.goal.com/nl/news/545/primera-division/2016/01/18/) ‘Preses Barcelona: “Next thing [they will say that] Neymar is involved in the murder of JFK”’ Vilarrubi points out the absurdity of the continuing allegations by saying that “before long” Neymar will be accused of murdering J. F. Kennedy. As such, and just like the speaker in (21), he is using a Reductio ad Absurdum (RAA) by presenting a clearly absurd standpoint as somehow analogous to the standpoint of his antagonist(s). Rather than acts and ensuing events, the argumentation concerns knowledge and opinions and is based on the relation of analogy in (23).13 (23) X is absurd Because: X is comparable to Y And: Y is absurd What is special about this third type, as compared to the first two, is that these utterances do not present future events that can be prevented by some course of 13 See

Jansen (2007) on distinguishing between arguments from consequences and RAA.

196

R. Boogaart

action in the present. In the preceding section, use of the temporal adverb straks in such cases was explained by postulating an implicit “saying event” in the near future. Thus, (21) and (22) can be paraphrased as “Next (thing) you/they will say that x” and, indeed, the explicit pattern itself is often used to advance this kind of argument. (This equally holds for English next (thing) you’ll say that x.) An internet search makes clear that the most frequently used verbs in the explicit variant are zeggen (‘say’), beweren (‘claim’), and vertellen (‘tell’), often combined with the auxiliary gaan (‘to be going to’) and the scalar particle (ook) nog, resulting in the fixed pattern straks ga je (ook) nog [verb of saying] that (‘next you will [verb of saying] that’) which clearly functions as a linguistic indicator of analogy-based RAA. Strictly speaking, the argument is not about what the antagonist will “say next” but about the opinion (s)he holds at the moment of speaking,14 but by using straks and, optionally, a verb of saying, to present this kind of argument, it still has a temporal/causal flavor: the speaker suggests that Y, as absurd as it may be, does somehow temporally and causally follow from X; both when used in slippery slope argumentation and in RAA, apprehensive straks appeals to a scale going unavoidably from “bad” to “worse”. Just like it may be questioned for the slippery slope arguments in (18), (19), and (20), if there really is a causal chain connecting X and the extreme consequence Y, it may clearly be questioned for (21) and (22) if X is really comparable to the extreme standpoint Y or if they constitute instances of “false analogy”. In the next section, I will elaborate further on the fallaciousness of such argumentative moves and on the strategic potential of using an expressive construction as a presentational device.

12.4 Expressives Between Reasonableness and Effectiveness In the preceding section, three different argumentative contexts were distinguished in which apprehensive straks may be used. However, the construction is just one of many options available to the arguer to present these types of arguments and the question is how presenting such arguments by means of an expressive construction contributes to the rhetorical goal of the arguer. In addition to “topical selection” and “adaptation to audience demand”, the choice of particular presentational devices, such as apprehensive straks, is one of the three aspects of “strategic maneuvering” in the extended Pragma-Dialectical theory of argumentation (summed up by van Eemeren, 2010). While arguers are supposed to adhere to standards of reasonableness (their dialectical goal), they will choose and present their argumentative moves in the most effective way given their rhetorical goal. Maintaining this balance between being reasonable and being effective is what is known as “strategic maneuvering”. If the rhetorical pursuit for effectiveness is at the expense of the dialectical obligations, 14 Use

of a verb of saying in this type of argument, therefore, constitutes an instance of “fictive interaction” in the sense of Pascual (2014).

12 Expressives in Argumentation: The Case of Apprehensive …

197

formalized as ten rules for critical discussion, the maneuver is said to be “derailed” (i.e. fallacious).15 While investigating strategic potential is relevant for all presentational devices, it seems particularly interesting for the category of expressive utterances. In the preceding section, we saw that one of the typical uses of the straks-construction is to present a fallacious argument and one might therefore expect for the fallacy to be immediately recognized as such. More generally, an appeal to emotion is one of the standard examples of non-argumentation (e.g. van Eemeren & Grootendorst, 1992, pp. 132–135) and if the semantics of an expressive construction includes such an appeal, one may wonder if this does not contribute negatively to its effectiveness. In this section, I will claim that, at two different levels, the construction functions as an “immunization strategy” (van Eemeren & Grootendorst, 1992, p. 119), making it difficult to criticize the argumentative move. First, the expressive dimension reinforces the argument but is “beyond the speaker’s control” (Sect. 12.4.1). Second, the argument presented by means of the construction is typically unreasonable, but overtly so, thereby “flouting” rather than violating the rules for critical discussion (Sect. 12.4.2).

12.4.1 The Expressive Dimension The general rhetorical benefits of using emotion as a persuasive tool are well known: it may function as an intensifying device, reinforcing the argumentation by the speaker, who shows, or at least suggests, strong personal engagement with the issue at hand; in more formal activity types, such as political discourse, it may also be used as a way to bring in elements of everyday talk and appeal to a larger audience (Alcaide Lara et al., 2016, p. 155). Such general remarks on the function of emotion in argumentation are certainly relevant for apprehensive straks as a presentational device since by using an expressive construction the arguer is putting forward his argument with some force, which may contribute to its effectiveness. To this general function of emotion in argumentation, we will add some considerations pertaining specifically to the use of expressive constructions. In Sect. 12.2, a distinction was made between expressing an emotion and describing an emotion. While apprehensive straks adds an expressive dimension, utterances containing the construction are not exclusively expressive since they have descriptive content as well and may therefore easily be reconstructed as an argument in the form of an assertive. Therefore, the construction really takes up a middle position between pure expressives, as in (24), and utterances that describe an emotion as in (25).16 15 The notions of derailment and fallacy are not entirely synonymous since a strategic maneuver might also derail towards reasonableness, but for the use of a rhetorical stylistic device such as apprehensive straks this does not seem to be a possible scenario. 16 How to distinguish between expressive and descriptive utterances, and especially between expressive and descriptive aspects of one and the same utterance, is an issue of some debate in linguistics (Foolen, 2015, pp. 477–478). For analytical purposes I am making a distinction between the

198

R. Boogaart

(24) Omg! Shit! Fuck! Damn! Aargh! … (25) I fear that p The purely expressive “utterances” in (24) do not have descriptive content. They show the speaker’s emotion without necessarily being interactional, let alone argumentative. As such, they suggest that they are a direct reflection of what the speaker feels and their expression is unmediated by any conscious reflection. In contrast, (25) is an instance of telling rather than showing: in the complementation construction, the emotion itself has become an “object of conceptualization” (Verhagen, 2005, p. 7). Making explicit the specific nature of the emotion obviously requires some distance or reflection on the part of the speaker. On the continuum from expressive to descriptive utterances, apprehensive straks is somewhere in between (24) and (25) and its status as an “expressive-descriptive” construction contributes to its strategic potential. On the one hand, straks (nog) p! has descriptive content like (25) since a value for p has to be provided for the utterance to be grammatical at all. The construction constitutes an emotionally charged stylistic device for presenting an argument, but it still presents an argument. In the words of Krabbe and van Laar (2015, p. 1182), “[I]f the emotion merely functions to convey a proposition that is part of the proponent’s case, then a possible charge of nonargumentation by the opponent would not be sustainable”. On the other hand, the construction constitutes the expression of an emotion, as in (24), rather than a description, as in (25). As a result, use of the construction makes clear that the speaker is emotionally involved, but it is hard to put into words exactly what he is feeling (cf. Pott’s notion of “descriptive ineffability” in Sect. 12.2.1). The expressive dimension of the construction, as manifest especially in its prosodic properties, suggests that the feeling is beyond the speaker’s control. Using an expressive construction to present an argument, then, functions as an immunization technique since the arguer “cannot help feeling the way he does”. In response to an expressive utterance, questioning the way the speaker feels amounts to accusing him of faking the emotion, which would constitute an extremely face threatening act. Making the feeling explicit, as in (25), is more risky as an argumentative move since it lacks the dimension of “involuntary” expression. Moreover, by making the emotion itself part of the propositional content of the utterance, and presenting the argument explicitly as the speaker’s own feeling by means of a complementation construction (e.g. I fear that x) (cf. van Leeuwen, 2014, pp. 234–237), it more readily invites critical questions about the extent to which the emotion is justified or relevant for the standpoint under discussion. Using an expressive-descriptive construction such as apprehensive straks, where the emotion is “hidden” in grammatical features such as word order and prosody and merely a presentational device, enables the arguer to reinforce the argument without making the appeal to emotion all too obvious.

expressive dimension as a presentational choice (Sect. 12.4.1) and the contents of the argument itself (Sect. 12.4.2) but they are not independent since p in straks (nog) p! has to be quite serious to justify the strength of the emotion being expressed.

12 Expressives in Argumentation: The Case of Apprehensive …

199

12.4.2 Overt Unreasonableness While it may be difficult to criticize the emotion expressed by straks nog p!, one may of course still criticize the descriptive contents of the utterance, i.e. the argument put forward itself. In order to evaluate if an argument presented by means of apprehensive straks constitutes a derailed strategic maneuver, it is useful to take into account the three types of argumentative contexts distinguished in Sect. 12.3. More specifically, use of the construction in warnings (type 1) should be set apart from its use in slippery slope argumentation (type 2) and RAA (type 3).17 For the latter two types, it may be observed that the construction is typically used in seemingly fallacious arguments, which means that the critical questions pertaining to the specific argument scheme, concerning either the presumed causal connection between X and Y in (15) or the comparability of X and Y in (23), cannot be answered satisfactorily. This raises two questions. First, why is apprehensive straks a “linguistic indicator” of fallacious argumentation, but only in two of the three types? Second, how can using a linguistic device that is indicating a fallacy contribute to the effectiveness of an argument? The reason that an “emotional” argument will not easily derail in the illocutionary act of warning (type 1) is that warnings are per definition, as part of their felicity conditions, in the interest of the addressee and are meant to prevent concrete, sometimes even physical harm, as in (16) and (17). The emotion expressed by apprehensive straks is strong, but in the context of a warning this is justified by the need to alert the addressee to imminent danger and the way to avert it. The strength and evaluative orientation of the construction (see 12.2.2) is the same for all three types, but in the case of rather long-term consequences (type 2) or more abstract ideas (type 3) there is no clear motivation for using a construction as expressive as straks (nog) p! unless the consequences are extremely undesirable (type 2) or the standpoint to be attacked is really quite absurd (type 3). In other words, the contents of p in straks (nog) p! has to be very strong in order to “keep up with” the expressive strength of the construction. As a result, outside of contexts of concrete danger, apprehensive straks is often used for the rhetorical trope of hyperbole, thus in utterances that are clearly exaggerated, as is the case both in type 2 and type 3. Typical examples from the preceding section are repeated in (26) and (27). (26) Van wie is dat stuk gespuis? Straks heb ik Artis in me huis! ‘Who does that piece of scum belong to? Pretty soon I will have Artis [the Amsterdam Zoo] in my house!’ (27) Preses Barcelona: “Straks is Neymar nog betrokken bij dood JFK” (https://www.goal.com/nl/news/545/primera-division/2016/01/18/) ‘Preses Barcelona: “Next thing [they will say that] Neymar is involved in the murder of JFK”’ 17 The fact that type 1 is different from type 2 and type 3 is also suggested by English that uses next

(thing) and pretty soon in the latter two types but not in the first one (cf. note 1). In Dutch, use of the scalar particle nog (translatable in this context as ‘even’) emphasizes the “hyperbolic” nature of the construction in type 2 and type 3, but it is not excluded in type 1 either (Boogaart, 2009, p. 177).

200

R. Boogaart

In order to understand why such moves are strategic, it is useful to draw an analogy with the “flouting” of conversational maxims as discussed by Grice (1975, p. 49). As is well known, Grice’s maxims are not always adhered to in conversation. If the violation is covert, we are dealing with non-cooperative, deceptive behavior. For instance, in the case of the maxim of Quality, that tells the speaker to be truthful, this means that the speaker is lying. However, maxims may also be overtly violated, or rather “flouted”, in order to trigger implicatures. In the case of hyperbole, the literal meaning of “what is said” is blatantly false and yet the speaker will not be accused of lying. Rather, since the violation is overt, the hearer may conclude that the speaker really meant something else than what he literally said.18 Crucially, in such cases, the speaker is not committed to the literal meaning of what he said, but only to its implicatures, as Grice (1975, p. 53) also assumes for instances of irony and metaphor. Analogous to the way the Gricean maxim of Quality can be flouted by being overtly untruthful, the rules for critical discussion may be flouted by being “overtly unreasonable”. Some presentational devices seem to do exactly this and the use of straks in type 2 and type 3 may well be one of them. Other examples of the phenomenon in Dutch are (28) and (29). (28) Wij van X adviseren X ‘We [producing X] advice the use of X’ (29) Je bent zelf een X ‘You are an X yourself’ The slogan in (28) was used for the first time in the full form Wij van w.c.-eend adviseren w.c.-eend (‘On behalf of w.c.-eend we advise the use of w.c.-eend’) in a tv-commercial in 1989 to promote the use of wc-eend, a toilet cleaning product. These days, the pattern is still used, as an accusation of the argumentum ad verecundiam fallacy, either with w.c.-eend itself or another company or product figuring as X (Sanders, 2019). However, the original slogan was obviously meant as joke and accusing the company of having committed a fallacy would only make sense at the level of literal content (Boogaart, 2013, p. 291). The pattern in (29) was described as a kind of grammaticalized tu quoque fallacy by Audring (2013), who shows that users of the construction are quite creative in their choice of X, specifically by using terms that, outside of the construction, cannot be applied to people or do not express a negative evaluation. An example is given in (30), used on Twitter to air frustration in response to the Dutch railways adjusting the timetable of the trains in view of the weather forecast. (30) Je bent zelf een aangepaste dienstregeling. ‘You are an adjusted timetable yourself’ 18 Dynel (2018) offers a book-length discussion of overt versus covert untruthfulness, associating the first one mainly with humor and the second one with deception. As for humor, it may be noted that many instances of apprehensive straks in type 2 and type 3 are (also) meant to be funny, which provides arguers with yet another escape route if they are accused of being unreasonable (“I was just kidding of course”).

12 Expressives in Argumentation: The Case of Apprehensive …

201

The point of these examples is that they are overtly violating a discussion rule, i.e. the Argument Scheme Rule in (28) and the Freedom Rule in (29). As a result, accusing these speakers of having committed a fallacy will be not very effective since they clearly cannot and will not deny that they did and in fact the fallacy was meant to be recognized by the addressee. In a similar way, arguers can use straks (nog) p! to flout the Argument Scheme Rule by incorrectly applying causal or analogical argumentation, but overtly so.19 This certainly counts as an immunization technique since it is hard to accuse someone of having committed a fallacy if they just did so explicitly. In (26), pointing out to Dirkie’s mother that keeping one dog will not necessarily, in the end, lead to the Amsterdam Zoo living in their house is stating the obvious, and not something that Dirkie’s mother will deny. In response to (27), one may argue that the idea that Neymar murdered JFK is absurd, but that would be confirming the point of Vilarrubi’s argument. In fact, when asking a critical question pertaining to the literal content of such utterances, arguers seems to be violating the Usage Rule themselves by being uncooperative and “misusing implicitness” (van Eemeren & Snoeck Henkemans, 2017, p. 167). Both Dirkie’s mother and Vilarubbi will readily accept that the literal meaning of what they said was unreasonable, but this would not really hurt their case. While they can easily deny commitment to the literal content, as they were “obviously exaggerating”, they did say what they said and thereby conjured up the extreme end of a negative scale while remaining vague about the weaker proposition they do commit to (cf. Snoeck Henkemans, 2017, p. 272).20 In this respect, the strategic potential of using overt unreasonableness may be compared to that of using praeteritio (Snoeck Henkemans, 2009) since both boil down to “not meaning to say x” and yet “saying x” at the same time. Compared to the existing literature on strategic maneuvering, analyzing the arguments presented by means of apprehensive straks as instances of overt unreasonableness, presents an interestingly different case. While strategic maneuvers are usually analyzed as ways to cover up or disguise fallacies (e.g. Garssen, 2009; van Eemeren, Garssen, & Meuffels, 2015), this case study suggests that it may sometimes be strategic to commit them overtly.

12.5 Conclusion This chapter analyzed the argumentative use of an expressive linguistic pattern from Dutch, consisting of a temporal adverb indicating “immediate future” in clause-initial 19 The straks-construction is a somewhat weaker linguistic indicator of overtly

fallacious argumentation than the two constructions in (28) and (29) since it is not by definition used to present a fallacy, as is evidenced by its use in warnings. 20 According to Claridge (2011, p. 12), a speaker using hyperbole “is committed to the deeper emotional and interactional, thus social truth of the statement” but this obviously makes it hard to determine what exactly the arguer carries the burden of proof for.

202

R. Boogaart

position, usually straks (‘shortly’), and specific prosodic properties. It was shown that the pattern constitutes a construction in the sense of Construction Grammar since its formal properties systematically correspond with a specific reading, in which the temporal meaning of the adverb may be backgrounded: straks nog p! expresses the speaker’s fear that p might be true. As such, the construction is naturally used in argumentative discourse for negative pragmatic argumentation, including warnings, but also for (a subtype of) Reductio ad Absurdum. From the viewpoint of strategic maneuvering between dialectical and rhetorial goals, the use of expressive linguistic patterns may easily “derail” since they involve an appeal to emotion and they are typically used to present exaggerated claims, resulting from argument schemes being incorrectly applied. The reason that the use of an expresssive construction to present an argument may still be effective is that it is difficult to criticize such an argumentative move. It may therefore be considered an immunization technique, in two respects. First, the exact nature of the emotion is not explicit and hard to pin down but the expressive dimension of the construction suggests that it is a direct reflection of how the speaker feels and that it is beyond the speaker’s control. Second, in accordance with the strength of the emotion being expressed, the descriptive content of the utterance will often be hyperbolic. This rhetorical trope enables the arguer to deny commitment to the literal meaning without making clear the weaker proposition he is committed to. It was argued for the apprehensive straks-construction that its strategic potential may be in flouting rather than violating the rules for critical discussion: one of the ways to prevent being accused of a fallacy is committing it overtly.

References Alcaide Lara, E., Carranza Márquez, A, & Fuentes Rodríguez, C. (2016). Emotional argumentation in political discourse. In C. Fuentes Rodríguez & G. Álvarez Benito (Eds.), A genderbased approach to parliamentary discourse. The Andalusian parliament (pp. 129–159). Amsterdam/Philadelphia: John Benjamins. Angelo, D., & Schultze-Berndt, E. (2016). Beware bambai—lest it be apprehensive. In F. Meakins & C. O’Shannessy (Eds.), Loss and renewal: Australian languages in contact (pp. 255–296). Boston/Berlin: De Gruyter. Anscombre, J. C., & Ducrot, O. (1983). L’argumentation dans la langue. Bruxelles: P. Mardaga. Audring, J. (2013). “Je bent zelf een universele afstandsbediening”. Hoe een schoolpleinzinnetje Twitter veroverde (“You are a universal remote control yourself”. How a school yard phrase conquered Twitter). Onze Taal, 82, 88–89. Boogaart, R. (2009). Een retorische straks-constructie (A rhetorical straks-construction). In R. Boogaart, J. Lalleman, M. Mooijaart, & M. van der Wal (Eds.), Woorden wisselen. Voor Ariane van Santen bij haar afscheid van de Leidse universiteit (pp. 167–182). Leiden: Stichting Neerlandistiek Leiden. Boogaart, R. (2013). Strategische manoeuvres met sterke drank: redelijk effectief? (Strategic manoeuvres with hard liquor: Reasonably effective?). In Th A J M Janssen & T. van Strien (Eds.), Neerlandistiek in Beeld (pp. 283–292). Amsterdam/Münster: Stichting Neerlandistiek VU/Nodus Publikationen.

12 Expressives in Argumentation: The Case of Apprehensive …

203

Boogaart, R. (2017). Straks heeft het Nederlands nog een apprehensive! (Next thing Dutch will have an apprehensive!). VakTaal, 30, 24. Boogaart, R., & Reuneker, A. (2017). Intersubjectivity and grammar. In B. Dancygier (Ed.), The Cambridge handbook of cognitive linguistics (pp. 188–206). Cambridge: Cambridge University Press. Chafe, W. (1988). Punctuation and the prosody of written language. Written Communication, 5, 395–426. Claridge, C. (2011). Hyperbole in English. A corpus-based study of exaggeration. Cambridge: Cambridge University Press. Corver, N. (2016). Emotion in the build of Dutch: Deviation, augmentation and duplication. Tijdschrift voor Nederlandse Taal- en Letterkunde, 132, 232–275. Ducrot, O. (2009). Slovenian lectures. Introduction into argumentative semantics. Ljubljana: Pedagoški inštitut. Dynel, M. (2018). Irony, deception and humour. Seeking the truth about over and covert untruthfulness. Boston/Berlin: Walter de Gruyter. Eemeren, F. H., Garssen, B., & Meuffels, B. (2015). The disguised ad baculum fallacy empirically investigated. Strategic maneuvering with threats. In F. H. van Eemeren, (Ed.), Reasonableness and effectiveness in argumentative discourse (pp. 815–826). Heidelberg: Springer. Fahnestock, J. (2011). Rhetorical style. The uses of language in persuasion. Oxford: Oxford University Press. Foolen, A. (2015). Expressives. In N. Riemer (Ed.), The Routledge handbook of semantics (pp. 473–490). London: Routledge. Garssen, B. (2009). Ad Hominem in disguise: Strategic Manoeuvring with direct personal attacks. Argumentation and Advocacy, 45, 207–213. Grice, H. P. (1975). Logic and conversation. In P. Cole, & J. L. Morgan (Eds.), Syntax and semantics, Vol. 3: Speech acts (pp. 41–58). New York: Academic Press. Hilpert, M. (2014). Construction Grammar and its application to English. Edinburgh: Edinburgh University Press. Jansen, H. (2007). Refuting a standpoint by appealing to its outcomes: Reductio ad Absurdum vs. Argument from Consequences. Informal Logic, 27, 249–266. Jansen, H. (2017). “You think that says a lot, but really it says nothing”. An argumentative and linguistic account of an idiomatic expression functioning as a presentational device. Argumentation, 31, 615–640. Krabbe, E., & van Laar, J. A. (2015). That’s no argument! The dialectic of non-argumentation. Synthese, 192, 1173–1197. Lichtenberk, F. (1995). Apprehensional epistemics. In J. L. Bybee & S. Fleischman (Eds.), Modality in grammar and discourse (pp. 293–328). Amsterdam/Philadelphia: John Benjamins. Pascual, E. (2014). Fictive interaction: The conversation frame in thought, language, and discourse. Amsterdam; Philadelphia: John Benjamins. Potts, Ch. (2007). The expressive dimension. Theoretical Linguistics, 33, 165–198. Sanders, E. (2019, March 12). Wij van WC-eend. NRC Handelsblad. https://www.nrc.nl/nieuws/ 2019/03/20/wij-van-wc-eend-a3953808. Searle, J. R. (1976). A classification of illocutionary acts. Language in Society, 5, 1–23. Snoeck Henkemans, A. F. (2009). Manoeuvring strategically with ‘praeteritio’. Argumentation, 23, 339–350. Snoeck Henkemans, A. F. (2017). Strategic manoeuvring with hyperbole in political debate. In F. H. van Eemeren & W. Peng (Eds.), Contextualizing pragma-dialectics (pp. 269–280). Amsterdam/Philadelphia: John Benjamins. Sweetser, E. (1990). From etymology to pragmatics. Cambridge: Cambridge University Press. Tomasello, M. (2003). Constructing a language: A usage-based theory of language acquisition. Cambridge, Mass./London: Harvard University Press. van Eemeren, F. H. (2010). Strategic maneuvering in argumentative discourse. Amsterdam/Philadelphia: John Benjmains.

204

R. Boogaart

van Eemeren, F. H., & Grootendorst, R. (1992). Argumentation, communication, and Fallacies: A pragma-dialectical perspective. Hillsdale, New Jersey: Lawrence Erlbaum Associates. van Eemeren, F. H., Houtlosser, P., & Snoeck Henkemans, F. (2005). Argumentatieve indicatoren in het Nederlands. Een pragma-dialectische studie. Amsterdam: Rozenberg Publishers. van Eemeren, F. H, & Snoeck Henkemans, A. F. (2017). Argumentation. Analysis and evaluation, 2nd edn. New York, London: Routledge. van Ginkel, A. (2018). Wie weet/Voor(dat) je het weet. Een onderzoek naar de sturende kracht van de constructies wie weet en voor(dat) je het weet (Wie weet/Voor(dat) je het weet. An investigation of the rhetorical orientation of the constructions wie weet (‘who knows’) and voor(dat) je het weet (‘before you know it’)) (MA thesis, Leiden University). van Haaften, T. & van Leeuwen, M. (2018). Strategic maneuvering with presentational devices: A systematic approach. In S. Oswald, & D. Maillat (Eds.), Argumentation and inference: Proceedings of the 2nd European Conference on Argumentation, Fribourg 2017, Volume II (pp. 873–886). London: College Publications. van Leeuwen, M. (2014). Systematic stylistic analysis. The use of a linguistic checklist. In B. Kaal, I. Max, & A. van Elfrinkhof (Eds.), From text to political positions. Text analysis across disciplines (pp. 225–244). Amsterdam: John Benjamins. Verhagen, A. (2000). “The girl that promised to become something”: An exploration into diachronic subjectification in Dutch. In T. F. Shannon & J. P. Snapper (Eds.), The Berkeley Conference on Dutch Linguistics 1997: The Dutch Language at the Millennium (pp. 197–208). Lanham, MD: University Press of America. Verhagen, A. (2005). Constructions of intersubjectivity. Discourse, syntax, and cognition. Oxford: Oxford University Press. Walton, D. (1992). Slippery slope arguments. Oxford: Clarendon Press.

Chapter 13

Argumentative Abduction in the Interpretation Process: A Pragma-Dialectical Study of an Ironic Utterance Antonio Duarte

13.1 Introduction Abductive reasoning is a form of everyday reasoning that is necessary and inevitable in our day-to-day lives, to the point that it is impossible for us to understand our peers without using it. In many instances, a correct interpretation [i.e. the sense, the semantic value of utterances in situations (Raccah, 2014)] has to be unraveled by evaluating a set of hypotheses that are inferred by the listener according to certain hints provided by the speaker. In such dialectical exchange, the interlocutors have to thoroughly examine the context in order to arrive at successful communication. Finally, “argumentative” abduction has to be performed based on contextual elements, i.e., from the listener’s knowledge at that point in the exchange. To evaluate the rapid, almost instantaneous, new hypothesis arrived at by abduction that leads to understanding between interlocutors in a reasoned way, here I focus on the analysis of an ironic utterance from a pragma-dialectical perspective (van Eemeren & Grootendorst, 1984). As the model of a critical discussion serves as a heuristic tool in the process of analytic reconstruction and as an evaluative tool in the process of critical assessment, my analysis could reveal some essential argumentative elements involved in this exchange. Moreover, irony as a rhetorical device mainly works since there is a “secret” message that has to be discovered. As Tindale states: Rhetorical devices like allusion and irony are effective because they operate in cognitive environments, which they modify and expand. The argumentation is received in ways that balance the objective and the personal. Moreover, the audience actively contributes by filling

A. Duarte (B) Department of Logic and Theoretical Philosophy, Universidad Complutense de Madrid, Madrid, Spain e-mail: [email protected] © Springer Nature Switzerland AG 2020 F. H. van Eemeren and B. Garssen (eds.), From Argument Schemes to Argumentative Relations in the Wild, Argumentation Library 35, https://doi.org/10.1007/978-3-030-28367-4_13

205

206

A. Duarte

the gap or seeing the association. In doing this, they are expressly addressed by the argumentation, and stand out from other recipients of the material who are not personally affected. (Tindale, 2015, p. 210)

Therefore, according to extended pragma-dialectics (van Eemeren & Houtlosser, 2000) which accounts for the rhetorical dimension in a critical discussion, these kinds of assertions could be understood as legitimate strategic maneuvering (Martínez Fabregat, 2014). Interestingly, here, the strategic maneuvering is based on the speaker’s assessment of the abductive ability of the listener. The associated pragma-dialectical fallacy which is strategically committed through violation of the dialectical rules, may act as an abductive trigger and may help us to discover certain intentions of our interlocutor, at least as far as the interpretation of the utterance is concerned. Therefore, here, abduction also plays an important role in achieving the dialectical purpose. This analysis reveals that these kinds of hypotheses that are drawn up in the interpretation process could be evaluated in an argumentative way, as answering certain critical questions regarding the theoretical norms of extended pragma-dialectics.

13.2 Abduction First, I will clarify the term “abduction” and identify its essential characteristics. As well as providing an operational delimitation of this term, my main purpose in this section is to show that this type of reasoning has an implicit “argumentative” and “methodological” dimension. In a broad sense, abduction is reasoning in which a new idea is introduced (the conclusion of which is a hypothesis) that is both tentative and relative to a given context. In the last period in his thought (after 1900), Peirce gave his final definition of abduction. When we are confronted by amazing events, we search for an explanation: The explanation must be such a proposition as would lead to the prediction of the observed facts, either as necessary consequences or at least as very probable under the circumstances. A hypothesis then, has to be adopted which is likely in itself and renders the facts likely. This step of adopting a hypothesis as being suggested by the facts, is what I call abduction (Peirce, CP 7.202, 1901).

In Fann’s words: “Any synthetic proposition, whether it is a nonobservable entity or a generalization (so-called), in so far as it is for the first time entertained as possibly true, it is an hypothesis arrived at by abduction” (Fann, 1970, pp. 33–34). The logical form of abduction would be: The surprising fact, C, is observed; But if A were true, C would be a matter of course; Hence, there is reason to suspect that A is true (Peirce, CP 5.189, 1903). The conclusion is a hypothesis based on what is known at a given point in the investigation; this conclusion will be evaluated and defined in the light of new discoveries. Due to the provisional nature of the hypothesis, it is necessary to fill certain

13 Argumentative Abduction in the Interpretation Process …

207

gaps with new data or information. In this way, abductive reasoning is an “instinct” that depends on the conscious or unconscious perception of connections between different aspects of the world. However, abduction has an implicit logic, insofar as one can submit the hypotheses arrived at by abduction to criticism and justify one’s choice. Thus, abduction allows us to formulate a general prediction, but without any guarantee of a successful outcome. Moreover, abduction, as a prognostic method, offers the only hope of regulating our future conduct in a rational way. As West (2016) points out, abductive reasoning plays a key role in Peirce’s concept of “virtual habit”. In the processes of generating an ongoing event picture in one’s mind, or of representing episodes in the future within one’s thoughts, virtual habits (i.e., virtual attained beliefs) determine which newly conceived hypotheses are more viable or will produce a particular outcome. Virtual habits represent a “natural” or “unconscious” method which tracks early decision-making and implies a mental evaluation of plausible hypotheses arrived at by abduction. In the context of scientific research, such plausible hypotheses should be tested by following the three stages of scientific research, as laid out by Peirce in his later work. Therefore, abduction would be the first step in investigation. After adoption of the hypothesis, we then have to trace its experimental consequences; this would be the second step in investigation: deductive reasoning. The last step is to verify the hypothesis by comparing the predictions deduced from it with the results of the experiment: inductive reasoning. If they match, the hypothesis is verified. Thus, Peirce shows that the three stages of scientific research emerged from the three types of inference (Peirce, CP 8.209, 1905). However, it is reasonable to see both hypothesis generation and preliminary evaluation as essential elements of abduction1 also in a more general sense, in common abductions of our day-to-day lives, as well as in the case of virtual habits. In Paavola’s words: “Abduction was then seen from the methodological perspective rather than as an evidencing process.” (Paavola, 2004, p. 246). In some way, abduction always involves the three Peircean stages of scientific research. Let us consider an example of common abduction in everyday life. Peirce offers us an example, taken from his own experience: I once landed at a seaport in a Turkish province; and, as I was walking up to the house which I was to visit, I met a man upon horseback, surrounded by four horsemen holding a canopy over his head. As the governor of the province was the only personage I could think of who would be so greatly honored, I inferred that this was he. This was a hypothesis. (Peirce, CP 2.625, 1878)

Peirce adopted the general rule that only a prominent personality is surrounded by four men carrying a canopy; since the most important personality in the province 1 Paavola

(2004, p. 266) connects this to Peirce’s idea of three interpretants: “Abduction should take into account the actual suggestions produced (cf. dynamical interpretant), and the “final” estimation of the hypothesis as a hypothesis (cf. final interpretant), but also that phase where abductive conclusion is only potential and inherent in various information and clues concerning the subject area in question (cf. immediate interpretant). The dynamical and the final interpretants would be equivalent to abduction in the generative and in the preliminary evaluative sense, respectively.”

208

A. Duarte

was the governor, Peirce conjectured that he had bumped into the governor. It could have been some other very powerful man, or a local tradition, but with the limited data available, his hypothesis was a plausible solution. New data could have been collected, for example: Where was the supposed governor going? How did people who crossed his path behave? Answers to such questions could have suggested a better explanation; but with the knowledge available at the time, we can consider the hypothesis to be appropriate. In this simple example, we see how Peirce’s abduction is supported by deductive inferences and, although experimental justification has not been carried out, it is considered plausible thanks to “inductive knowledge”. As the foregoing example shows, abduction is usually presented in open circumstances, i.e., in situations where our knowledge is incomplete. This process is not only one of reasoning, but also of dialogue insofar as we are continually asking and answering questions. As Hintikka points out, there is a close affinity between Peircean abduction and the interrogative model of inquiry, especially through explanationseeking “why” questions: “Abductive ‘inferences’ must be constructed as answers to the inquirer’s explicit or (usually) tacit questions put to some definitive source of answers (information)” (Hintikka, 1998, p. 519). As a dialectical process, abduction progresses within specific dialogic frameworks. Thus, abductive reasoning, understood as the whole process of searching for hypotheses that explain and normalize an unusual event, can be treated as an exchange of arguments in a critical discussion aimed at achieving a specific goal. Although the aim of this dialectical process is the resolution of an enigma rather than the resolution of a difference of opinion, a dialectical approach to abduction enable us to assess whether the objective (enigma resolution) is achieved in a reasonable way in accordance with the contextual framework. Moreover, from a pragma-dialectical perspective, the rules that we have to observe when we are involved in a critical discussion (see e.g. van Eemeren & Grootendorst, 1984, 1987) also help us to perform sound abductions (not infallible, but well argued). It is worth noting that this dialectical approach to abduction is especially well suited to exploring and analyzing abductions generated in a context of scientific inquiry (understood in a broad sense), where it is mandatory that the conclusions be based on solid grounds and sound arguments. In many instances, the abductive process may last several years and implies a sort of reasoned argumentative exchange between, for example, “one’s hypotheses” and “objective data”; “the questions one is prompted to ask” and “the search for new evidence”; or, turning to a real critical discussion, between “researcher 1” and “researcher 2”. In situations that are closer to our everyday lives, as in many instances of the interpretation process, this dialectical model may well also be useful if we are interested in determining the role of abductive reasoning in them. This is because abduction always arises in contextualized situations and is far from a simple guessing instinct or the generation of decontextualized hypotheses.

13 Argumentative Abduction in the Interpretation Process …

209

13.3 Conversational Implicature and Irony To analyze “argumentative” abductions during the interpretation process using extended pragma-dialectics, an example of an ironic utterance will be presented. This section is therefore devoted to detailing some theoretical perspectives that address the particularities of this speech act. I briefly summarize Grice’s theory of conversational implicature and different proposals for pragmatic approaches to irony. In Logic and Conversation, Grice states: Our talk exchanges do not normally consist of a succession of disconnected remarks, and would not be rational if they did. They are characteristically, to some degree at least, cooperative efforts; and each participant recognizes in them, to some extent, a common purpose, or at least a mutually accepted direction. (Grice, 1975, p. 45)

Based on this idea of common purpose, Grice derives his Cooperative Principle: “Make your conversational contribution such as is required, at the stage at which it occurs, by the accepted purpose or direction of the talk exchange in which you are engaged” (Grice, 1975, p. 45). After this, Grice divides this maxim into several sub-maxims, following the Kantian categories: quantity, quality, relation, and manner. Clearly, many of our conversations are not guided by this principle. However, violation of the rules puts us on alert, and may give rise to an enigma. The rules can be violated in several ways. There may be conflicts between the maxims: for example, failing to provide relevant information because there is not enough evidence for it. There may also be conflict with other purposes: for example, someone who has valuable and pertinent information for the purposes of conversation may fail to fulfill a maxim for reasons of discretion, saying “I cannot say more, my lips are sealed.” However, there is another way to break a rule that does generate conversational implicature. This is when one of the participants clearly violates a maxim with their words, according to the literal meaning of the utterance, i.e. the semantic value of the language units (Raccah, 2014), and yet it is not possible to explain such a violation in similar ways to those presented above. Let us see an example given by Grice: A is writing a testimonial about a pupil who is a candidate for a philosophy job, and his letter reads as follows: ‘Dear Sir, Mr. X’s command of English is excellent, and his attendance at tutorials has been regular. Yours, etc.’ (Gloss: A cannot be opting out, since if he wished to be uncooperative, why write at all? He cannot be unable, through ignorance, to say more, since the man is his pupil; moreover, he knows that more information than this is wanted. He must, therefore, be wishing to impart information that he is reluctant to write down. This supposition is tenable only on the assumption that he thinks Mr. X is no good at philosophy. This, then, is what he is implicating.) (Grice, 1975, p. 52)

The implicature is obtained by interpreting the speaker’s words in such a way as to remove the conflict with the conversational maxims from their literal meaning and appraising other contextual elements that the speaker and audience share. The intention of the speaker is therefore to make the audience aware of this violation and want to seek an explanation or an alternative hypothesis that satisfies the contextual elements involved. In other words, the speaker poses a riddle for the audience. Using Grice’s theory, a wide range of deviations of intent from that expressed by the standard

210

A. Duarte

or ordinary use of language, as in the case of jokes, puns, metaphors, and ironic utterances, could be understood as conversational implicature generated by flouting, violating, opting out of, or infringing at least one of the maxims derived from the Cooperative Principle. Indeed, Grice provides the first modern pragmatic definition of irony. Traditional theories of irony assume that an ironist uses a figurative meaning that is opposite to the literal meaning of the utterance. Thus, a person who says “what lovely weather” on a rainy day is using the figurative meaning, “what terrible weather” (see e.g. Clark and Gerrig, 1984). Grice recognizes that just saying something that is obviously false is not sufficient for irony. According to Grice (1975, 1978), irony is an instance of speakers or writers flouting the quality maxim. The category of quality provides a super-maxim: “Try to make your contribution one that is true” (Grice, 1975, p. 46), which is clarified via two others: “1. Do not say what you believe to be false. 2. Do not say that for which you lack adequate evidence” (Grice, 1975, p. 46). Grice’s theory assumes that the ironist is pretending rather than using one proposition in order to express the contrary. Moreover, to work out the implicature, the interlocutor has first to address the literal meaning of the utterance. Therefore, Grice assumes that irony presupposes two-stage processing, which would involve the processing of a literal meaning, the rejection of this interpretation on pragmatic grounds, and a subsequent reinterpretation. In contrast to this, Sperber and Wilson (1981), based on some of the postulates of Grice’s theory, offered a mention theory of irony (further developed and embedded in their Relevance Theory; see e.g. Wilson & Sperber, 2004; Wilson, 2006). In that, speakers are being ironic when they mention, or echo, an earlier utterance, while dissociating themselves from the echoed opinion via accompanying ridicule or scorn. The central claim of relevance theory is that “the expectations of relevance raised by an utterance are precise and predictable enough to guide the hearer towards the speaker’s meaning” (Wilson & Sperber, 2004, p. 607). Thus, this mention/relevance theory claims that the processing of irony is not distinct from that of literal meaning and that ironic meaning is arrived at directly (one-stage processing). Tindale (2015, pp. 99–116) offers a broad discussion of these primary theories regarding irony. Following the claim of Saul (2002), who proposed that the two accounts are analyzing different aspects of what is said, from different perspectives, Tindale draws an interesting conclusion which reconciles the different views: Grice is primarily (but not solely) interested in how the speaker’s intentions convey meaning; the relevance theorists are interested in how audiences interpret meanings. Together, they give a fuller sense of how communication works from the different perspectives. The danger is that committing fully to either account results in an incomplete picture. (Tindale, 2015, p. 116)

Be that as it may, the one-stage/two-stage distinction has taken center stage in the debate over irony. Evidence supporting the direct-access view of irony comes from equal reading times of ironic and non-ironic utterances (see e.g. Gibbs, 1986; Gibbs & O’Brien, 1991). However, experiments carried out by Giora and collaborators report that utterances took longer to read in ironically than in literally biased contexts (see

13 Argumentative Abduction in the Interpretation Process …

211

e.g. Giora, 1997; Giora & Fein, 1999). These latter findings are therefore consistent with a two-stage view of irony comprehension and the authors derived a more general Principle of Salience which is based on the idea that more salient meanings take priority in interpretation. According to the graded salience hypothesis […] the factor determining initial activation is neither literality nor compatibility with context, but rather the salience of the verbal stimulus: Salient meanings of words and expressions should always be accessed and always first. A meaning of a word or an expression is salient if it is coded in the mental lexicon. […]. According to the graded salience hypothesis, then, the salient meaning of a word or an expression is accessed directly. When it is contextually compatible, no more processes are required. However, when a less salient meaning has to be activated to make sense of an utterance […] comprehension should involve an ordered access: The more salient, albeit inappropriate meaning should be processed initially, before the less salient, appropriate meaning can be retrieved. (Giora & Fein, 1999, pp. 242–243)

Although this graded salience hypothesis rejects the priority of literal meaning in favor of salient meanings and accounts for one-stage processing in many cases, it retains the Gricean two-stage model, since the processing of novel ironies requires backtracking and reinterpretation. Moreover, the definition of salience is a central issue when addressing its practical differentiation from literal meaning. As Attardo (2000) summarizes: Salience is defined as a function of 1. a word’s meaning conventionality (i.e. whether a word has that meaning by convention; in other words, if it is listed in the lexicon as having that meaning), 2. familiarity (e.g. freedom is more familiar than liberty), 3. frequency (i.e. more frequent meanings are more salient), and 4. ‘givenness status’ in a linguistic co(n)text (Giora, 1997: 185). (Attardo, 2000, p. 800)

The following micro story may clarify this concept: “‘You’re a monster’, she shouted. He nodded assent with what seemed to be his head” (Olgoso, 1999; my own translation from the original Spanish)2 . In this situation, the salient meaning of “monster” refers to a person who arouses disgust by wickedness, cruelty, etc. However, this micro story evokes terror because it refers to the less salient meaning of “monster” in this context: an ugly non-human creature that frightens people. With regard to irony, the salient meaning of many well-known ironic expressions will be directly the intended meaning as, for example, “what lovely weather” on a rainy day. However, in many instances of ironic utterances, especially novel ones, while the salient meaning corresponds exactly to the literal meaning, this is not the intended meaning and a two-stage processing is required. In the following, I am going to assume the two-stage approach proposed by Attardo (2000), who develops Grice’s theory of irony, including the concept of salience and the presumption of relevance: 2 “Eres

un monstruo”, le gritó ella. Él asintió con lo que parecía su cabeza.

212

A. Duarte

The theory of irony • the reconstruction of the intended meaning (value) of the irony is entirely inferential and abductive: it is totally indirect, no aspect of the meaning is given in the text, except the presumption of relevance (and not of quality, manner, or quantity); or, in other words, irony is a purely pragmatic phenomenon; • irony is essentially an inappropriate utterance which is nonetheless relevant to the context; • irony crucially involves a two-stage processing. The order in which the conflicting senses are accessed is (probably) determined by salience. (Attardo, 2000, p. 823)

To clarify these issues, I am going to introduce here the case that I will analyze in the next section. This case was proposed by García-Carpintero (2008, pp. 492–493) in regard to Grice’s theory of conversational implicature. Begoña, who has certain feminist leanings, is driving a car and I am accompanying her. We are following behind another vehicle. The vehicle in front makes all kinds of unfortunate maneuvers of the types which exasperate other drivers. Finally, Begoña has the chance to overtake the car in front; on doing so, we both look with morbid curiosity at the driver of the other vehicle, perhaps trying to find some unmistakable sign of incompetence. The driver turns out to be a woman. So Begoña says: “It had to be a white car!” (García-Carpintero, 2008, pp. 492–493; my own translation from the original Spanish). 3

Begoña’s statement is an inappropriate utterance which is nonetheless relevant to the context; therefore, it poses an enigma. Conventionally, the statement emphasizes that it is a white car that has made those unfortunate maneuvers. The expression “had to be” makes an association between this type of bad driving and white cars. Recalling Grice’s conversational maxims, we can see this statement violates several of them. Begoña lacks the appropriate data to make such a claim (category of quality: “Do not say that for which you lack adequate evidence.”), and the statement is inappropriate in the sense that it is difficult to establish an association between the color of a car and the driver’s inability (category of manner: “Avoid obscurity of expression./Avoid ambiguity.”). Through detection of these violations of the Cooperative Principle, we can see that Begoña’s assertion has all the appearance of the fallacy of poor generalization, i.e., it is an inference from a proven fact to an excessive generalization. The violation of the maxims cannot be easily explained within the context, thus, we find that the speaker has generated a conversational implicature: “the speaker thinks (and would expect the hearer to think that the speaker thinks) that it is within the competence of the hearer to work out, or grasp intuitively [the implicature]” (Grice, 1975, p. 50). As pointed out above, the reconstruction of the intended meaning of the irony is indirect: no aspect of the meaning is given in the text, except the presumption of relevance. For most listeners, the more salient meaning of “It had to be a white car!” is exactly “It had to be a white a car!”, so a less salient meaning has to be activated 3 Begoña,

quien tiene ciertas inclinaciones feministas, conduce el coche y yo la acompaño. Nos precede otro vehículo. El vehículo que nos precede realiza todo tipo de maniobras desafortunadas, de esas que exasperan a los otros conductores. Finalmente, Begoña encuentra la ocasión de adelantarle; al hacerlo, ambos miramos con curiosidad malsana al conductor del otro vehículo—quizás tratando de encontrar en su rostro signos inequívocos de incompetencia—. El conductor resulta ser una mujer. Entonces Begoña dice: “¡Vaya, un coche blanco tenía que ser!”.

13 Argumentative Abduction in the Interpretation Process …

213

to make sense of this statement. To arrive at that less salient meaning, argumentative abduction has to be performed based on contextual elements. As Tindale points out: Maxims require interpretation, especially when we are dealing with potential figures of speech. Such interpretations must involve not just a focus on what is said and how it is said, but also the relation of this to a set of beliefs and assumptions held in common by speaker and audience. (Tindale, 2015, p. 116)

13.4 Irony from Extended Pragma-Dialectics In this section, I discuss a case of an ironic utterance (Begoña’s case presented in Sect. 13.3) from a pragma-dialectical perspective. My analysis will guide us towards the argumentative elements involved in these statements. Moreover, the application of the theoretical norms of this perspective to utterances in which the intended meaning has to be unraveled may shed some light on assessing the interpretation process.

13.4.1 Irony: Standpoint or Conclusion? In order to analyze this case using extended pragma-dialectics, hereafter I consider Begoña’s exclamation as a contribution to a critical discussion: an implicit argument that involves a pattern of reasoning from premises. In this case we start from: this white car makes unfortunate maneuvers; leading to a conclusion: there is some connection between the color white and driving badly. Moreover, this expression occurs in a context where it fulfills a specific function in the communication process, aimed at a certain goal. Depending on the perspective one takes, the statement may be analyzed either as a conclusion or as a standpoint. In pragma-dialectical theory, the object of argumentation is referred to as the standpoint. From this perspective, advancing a standpoint is tantamount to performing an assertive which implies a commitment to the truth or correctness of the propositional content of the speech act performed. Felicity conditions for advancing a positive standpoint are: Identity Conditions

Propositional Content Condition 1. The propositional content of the standpoint consists of an expressed opinion O. 2. O consists of one or more utterances.

214

A. Duarte

Essential Condition Advancing a standpoint counts as taking responsibility for a positive position in respect to O, i.e., as assuming an obligation to defend a positive position in respect to O, if requested to do so. Correctness Conditions

Preparatory Condition 1. Speaker S believes that listener L does not (already, at face value, completely) accept O. 2. S believes that he can justify O for L with the help of arguments. Sincerity Condition 1. S believes that O is the case. 2. S has the intention to justify O for L with the help of arguments if requested to do so. (Houtlosser, 2001, p. 32)

Taking into account the literal meaning of the expression, Begoña’s exclamation carries specific commitments and responsibilities. Therefore, as a standpoint is not viewed as a psychological attitude or mental state, Begoña’s assertive speech act fulfills the felicity conditions of advancing a standpoint, including the Sincerity Condition. However, ironic utterances are usually presented as statements that do not need elucidation. Ironic speakers suggest that the standpoint has already been accepted by their antagonist, whereas its acceptability was the issue all along. Therefore, although ironic utterances are (ironic) standpoints, speakers present them as conclusions to trigger the listener’s abduction. At this point, the hearer recognizes that the ironic speaker has violated the dialectical rules. From the pragma-dialectical perspective, there are certain rules that we have to observe when we are involved in a critical discussion (see e.g. van Eemeren & Grootendorst, 1984, 1987). These are related to questions about relevance or the balance of the burden of proof. Violations of the discussion rules are said to frustrate the reasonable resolution of the difference of opinion and they are therefore considered as fallacies. According to pragma-dialectics, Begoña’s statement violates rule VIII of those proposed by van Eemeren & Grootendorst (1987, pp. 284–290), because the argumentation scheme is improperly applied (“the arguments used in a discursive text must be valid or capable of being validated by the explicitization of one or more unexpressed premises”). In particular, by justifying a general conclusion from an insufficient number of observations, this is a case of hasty generalization or secundum quid. This violation warns the listener that there is something wrong with referring to the literal meaning of the utterance. Nevertheless, violation of rule VIII is not enough for a standpoint to become a hasty conclusion. The important point is that

13 Argumentative Abduction in the Interpretation Process …

215

Begoña’s words, as well as most ironic utterances, violate rules I (“parties must not prevent each other from advancing or casting doubt on standpoints”) and II (“whoever advances a standpoint is obliged to defend it if asked to do so”) as a standpoint is declared as sacrosanct. By violation of these rules, Begoña evades and shifts the burden of proof. In this case, the expression “had to be” in the assertion “invites” the listener not to argue with the issue. As in the case of conversational implicature, detecting a pragma-dialectical fallacy warns us that we should review the words of our interlocutor to be sure that we have to address the literal interpretation, and not some other one. Therefore, the theory of irony from a pragma-dialectical perspective should also assume a twostage processing approach. Interestingly, ironical arguers rely on the listener’s abductive capacity to advance a standpoint. When the listener recognizes the intended meaning of the irony, the ironic utterance becomes a real standpoint which is different from that expressed by the literal words of the speech act. This is a risk for ironical arguers, “so they must believe it can have a particularly effective payoff, not achieved through more direct means, in order to judge the risk worthwhile” (Tindale, 2015, p. 207). Therefore, abductive reasoning plays an important role in achieving the dialectical purpose.

13.4.2 Irony as Strategic Maneuvering As we saw above, a speaker’s violation of the dialectical rules warns us about interpreting the utterance, as we have both a literal meaning that cannot be explained according to the dialectical context and a standpoint presented as a conclusion. Therefore, we can say that when we are faced with ironical arguers, the rules are “cleverly” violated to produce an effective move in the course of the argumentation. To be effective, irony requires the stability of language and norms such as the dialectical rules. This idea is also pointed out by Tindale: When effective, irony arrests an audience’s attention by forcing a contrast in their expectation. Norms associated with an idea, a concept, or a practice would dictate one response, and the ironic statement or argument provides a contrasting one. (Tindale, 2015, p. 208)

Therefore, here the ironical arguer breaks the rules as a rhetorical device to achieve an effective move in the critical discussion. To take this rhetorical dimension into consideration in the analysis of pragma-dialectical reconstruction, the standard version of the pragma-dialectical theory was extended by including the concept of “strategic maneuvering” (van Eemeren & Houtlosser, 2000). This concept takes into account rhetorical maneuvers in the analysis of argumentative discussion. Parties involved in a difference of opinion “maneuver strategically” to realize their dialectical and rhetorical aims simultaneously: this is a way of being persuasive while observing the standards of a critical discussion. In analyzing the strategic function of the maneuvering, the following parameters need to be considered: 1. the results that can be achieved by the maneuvering;

216

A. Duarte

2. the routes that can be taken to achieve these results; 3. the constraints imposed by the institutional context; 4. the commitments defining the argumentative situation. (van Eemeren & Houtlosser, 2009, p. 11)

From this perspective, fallacies are also understood as derailments of strategic maneuvering: when the rational exchange is revoked to achieve greater efficiency, i.e. “if a party allows his commitment to a critical exchange of argumentative moves to be overruled by the aim of persuading the opponent so that his moves are no longer in agreement with the critical norms” (van Eemeren & Houtlosser, 2009, pp. 13–14). From the point of view of the speaker’s speech act, if we consider Begoña’s literal words as strategic maneuvering, the desire for efficiency (“had to be”) removes any kind of rationality from the issue. From the literal meaning of the words, Begoña has committed a fallacy: some rules of the critical discussion have been violated or the strategic maneuvering has derailed. Here, it is also worth noting that an ironic utterance can be considered as legitimate strategic maneuvering only if the audience infers the intended meaning of the expression. The strategic maneuvering is based, again, on the speaker’s assessment of the abductive ability of the interlocutor.

13.4.3 Interpreting Irony by Argumentative Abduction We saw above how we can assess some elements of a speaker’s ironic arguments using extended pragma-dialectics. I concluded that the abductive reasoning of the listener becomes essential for the speaker to achieve their argumentative goals. The abduction will be inferred according to the contextual elements. Therefore, in order to examine the situation thoroughly, the listener has to follow a method to arrive at the hypothesis which leads to success in the communication process. Although audiences follow this method in a rather “unconscious” way, we can derive some tools to assess the value of the listener’s abductive reasoning in the dialectical exchange. To clarify how abductive hypotheses are arrived at, I am first going to consider a different contextual situation from that expressed in the Begoña example, but concerning Begoña’s ironic utterance. Let us suppose that a third driver saw what happened and heard Begoña’s words, but without knowing the sex of the driver of the white car. Certainly, this new actor would have considered Begoña’s exclamation to be strange and misleading. How could these enigmatic words be explained, considering the few contextual data at hand? The third driver might have guessed at various different scenarios, for example, “the girl in that car probably suffered some trauma caused by a white car” or “she has a superstitious prejudice against the color white.” Notice that these attempts at an explanation cannot be confirmed by the driver because of the lack of pragmatic knowledge. This driver could not provide a plausible explanation to work out the implicature. Figure 13.1 shows an argumentation scheme to the best explanation (see e.g. Macagno & Walton, 2013) which summarizes the third driver’s attempts to interpret Begoña’s words. The scheme refers to

13 Argumentative Abduction in the Interpretation Process …

Fig. 13.1 Argumentation scheme to the best explanation for the case of the third driver

217

218

A. Duarte

the “method” one follows to arrive at a plausible hypothesis. The actor could reach a conclusion by adopting an explanation that is more difficult to refute based on their presumptions or knowledge. As in the example of an everyday abduction which appears in Sect. 13.2, the way one reaches the conclusion (abduction) is a rather instantaneous sort of example of the three Peircean stages of scientific research. The surprising fact (Begoña’s words) is explained by tentative hypotheses (explanations) which are supported by deductive inferences derived from the hypothesis and by inductive knowledge (presumptions and knowledge). In this case, the lack of a preferred hypothesis of the third driver demonstrates clearly how the speaker relies on the abductive ability of her interlocutor to achieve her dialectical purpose. This abductive capacity is based on different aspects: the interlocutors’ knowledge of the common dialogic framework and the listener’s ability to make connections between apparently unrelated phenomena. When interlocutors share a broader base of pragmatic knowledge, they have increased abductive capacity. This is translated into both an ironic utterance that is well-suited to the situation according to the speaker’s knowledge of the audience, and also better abduction to arrive at successful communication on the part of the audience. As opposed to the third driver, Begoña and I are friends and I am accompanying her. I can therefore suggest a far more plausible explanation, because I know two key contextual facts: 1. The driver of the white car is a woman. 2. Begoña is against the kind of prejudice that assumes that women are inferior to men in certain skills such as driving. In situations such as the one described, many men have been known to say: “It had to be a woman!” to support the hypothesis that women are worse drivers than men. Due to the second key fact I have to hand, I know that Begoña does not believe such things; and she knows, in turn, that I know this. Based on the assumed abductive competence of her audience, Begoña is saying: “Do not make use of spurious generalizations such as ‘there is a causal relationship between being female and driving badly’.” Begoña’s persuasive purpose is legitimate and, at this point, the hearer has recognized Begoña’s real standpoint. In addition, the hearer also recognizes the strategic function of Begoña’s maneuver: (a) by violation of rule VIII (see Sect. 13.4.1), the hearer realizes that the real standpoint has been presented by using an absurd argument which has the same form as what Begoña wants to denounce, and (b) by violation of rules I and II (see Sect. 13.4.1), Begoña invites the hearer to make the necessary contribution to appreciate her standpoint: listeners have to complete the reasoning, to see it for themselves. Therefore, the effectiveness of irony can be explained in terms of extended pragma-dialectics. Here, the speaker’s ironic utterance is understood as strategic maneuvering in the discussion that acts as an abductive trigger (Aliseda, 2006). It consists of violating the rules of critical discussion. The hearer notices that the dialogic rules have been broken and then adopts a hypothesis through which to arrive at a proper interpretation. Interestingly, in these cases, the efficiency of the strategic maneuvering is mainly based on the interlocutors’ abductive abilities. In the case of the third driver, this

13 Argumentative Abduction in the Interpretation Process …

219

strategic maneuvering is derailed due to the listener’s abduction failing: there is no clear explanation for the violation of the dialectical rules that leads to successful communication, i.e. Begoña has committed a pragma-dialectical fallacy. It is worth noting that we could imagine a wide range of hypotheses which would allow the implicature to be worked out (imagining different situations, dialogic contexts, and listener’s abductive capacities). As we have seen in the example above, the hearer in each context could reach a conclusion by adopting an explanation that is more difficult to refute based on their presumptions or knowledge. Although a conclusion, an “argumentative” and “methodological” abduction, is reached quickly, here I can offer some tools with which to evaluate this hypothesis in context. As well as reconstructing a proper argumentation scheme to assess the “method” followed to reach a conclusion, we may form some critical questions which address the theoretical norms of extended pragma-dialectics: 1. Does this hypothesis explain the speaker’s violation of the dialectical rules? 2. Does this hypothesis address the speaker’s desire for efficiency? In the context of the third driver, neither the first question nor the second can be answered in the affirmative. Although the alternative hypotheses arrived at try to explain Begoña’s words from a literal point of view (due to little pragmatic knowledge), there can be no satisfactory explanation for the speaker’s violation of rules I and II. In addition, the speaker’s desire for efficiency is not fully accounted for by adopting either of those hypotheses. Therefore, the third driver should realize that there is a failure of the required abduction to arrive at a correct interpretation. Moreover, only after answering these questions in the negative, could a listener who possesses broader pragmatic knowledge opt for an alternative hypothesis: the speaker has committed a fallacy. In contrast, when a hypothesis arrived at by abduction intended to recover the intended meaning of some strange words allows us to respond to one of these two questions in the affirmative, we may be in a position to discover the speaker’s standpoint by removing the conflict between the literal meaning of the utterances and the aims of the dialogue. To sum up, in order to arrive at successful communication, listeners have to search for hypotheses that provide affirmative answers to the proposed critical questions.

13.5 Conclusion Surprising facts trigger our abductive reasoning. However, abduction is wider than a “simple” inference insofar as it possesses an implicit methodological and argumentative dimension. For this reason, it should be possible to evaluate all conclusions derived from this type of reasoning. However, it is difficult to assess the argumentative dimension of abduction in instances where we infer new hypotheses rapidly, almost instantaneously, as those required in many cases of the interpretation process that lead to understanding between interlocutors. Here, to shed some light on this

220

A. Duarte

type of abduction, I have analyzed an ironic utterance from a pragma-dialectical perspective. By doing so, the theoretical norms of this perspective have revealed some argumentative elements which can help us to evaluate inferred hypotheses in context during the interpretation process. We have seen that, at first glance, after an ironic utterance is pronounced in the course of a critical discussion, we detect a pragma-dialectical fallacy due to the violation of some of the dialectical rules. Usually, in these statements, a standpoint is presented as a conclusion, thereby flouting pragma-dialectical rules I and II. However, the assertion could be legitimate strategic maneuvering which is based on the speaker’s assessment of the abductive ability of the interlocutor. The speaker thus evades the burden of proof and shifts it onto the other party in order to trigger the listener’s abduction. Hence, detecting a pragma-dialectical fallacy may act as an abductive trigger and can help us to discover certain intentions of our interlocutor, at least as far as the interpretation of the utterance is concerned. To correctly unravel the mysterious words, the listener has to propose an “argumentative” abduction from the contextual elements that would remove the conflict between the salient meaning of the utterance and the aims of the dialogue. This pragma-dialectical analysis assumes a two-stage processing approach to the theory of irony. Moreover, in deviations of intent from that expressed by the standard or ordinary use of language, the hearer’s inferred hypotheses about the proper interpretation could be evaluated in context by answering certain critical questions which address the theoretical norms of extended pragma-dialectics: (1) Does this hypothesis explain the speaker’s violation of the dialectical rules? (2) Does this hypothesis address the speaker’s desire for efficiency? When neither the first question nor the second question can be answered in the affirmative, the hearer could opt for alternative hypotheses that do not attempt to unravel the speaker’s enigmatic words. These alternative hypotheses would be: (a) the abduction required to arrive at a correct interpretation has not been reached due to limited contextual data; or, for listeners who are confident in their abductive abilities, (b) the speaker has committed a fallacy. We most certainly do not always want to evaluate the hypotheses arrived at by abduction, but if we do want to, it should prove extremely useful to follow the norms of the dialectical approach to argumentation.

References Aliseda, A. (2006). Abductive reasoning. Logical investigations into discovery and explanation. Dordrecht: Springer. Attardo, S. (2000). Irony as relevant inappropriateness. Journal of Pragmatics, 32, 793–826. Clark, H., & Gerrig, J. (1984). On the pretense theory of irony. Journal of Experimental Psychology: General, 113(1), 121–126. Fann, K. T. (1970). Peirce’s theory of abduction. The Hague: Martinus Nijhoff. García-Carpintero, M. (2008). Las palabras, las ideas y las cosas. Una presentación de la filosofía del lenguaje [Words, ideas and things. An introduction to philosophy of language]. Barcelona: Editorial Ariel.

13 Argumentative Abduction in the Interpretation Process …

221

Gibbs, R. W. (1986). On the psycholinguistics of sarcasm. Journal of Experimental Psychology, 115(1), 3–15. Gibbs, R. W., & O’Brien, J. E. (1991). Psychological aspects of irony understanding. Journal of Pragmatics, 16(6), 523–530. Giora, R. (1997). Understanding figurative and literal language: The graded salience hypothesis. Cognitive Linguistics, 8(3), 183–206. Giora, R., & Fein, O. (1999). Irony: Context and salience. Metaphor and Symbol, 14, 241–257. Grice, H. P. (1975). Logic and conversation. In P. Cole & J. Morgan (Eds.), Syntax and semantics 3: Speech acts (pp. 41–58). New York: Academic Press. Grice, H. P. (1978). Further notes on logic and conversation. In P. Cole (Ed.), Syntax and semantics (Vol. 9, pp. 113–117). New York: Academic Press. Hintikka, J. (1998). What is abduction? The fundamental problem of contemporary epistemology. Transactions of the Charles S. Peirce Society: A Quarterly Journal in American Philosophy, 34(3), 503–534. Houtlosser, P. (2001). Points of view. In F. van Eemeren (Ed.), Crucial concepts in argumentation theory (pp. 27–50). Amsterdam: Sic Sat, International Centre for the Study of Argumentation. Macagno, F., & Walton, D. (2013). Implicatures as forms of argument. In A. Capone, F. Lo Piparo, & M. Carapezza (Eds.), Perspectives on pragmatics and philosophy (pp. 203–224). Berlin/New York: Springer. Martínez Fabregat, S. (2014). La ironía como estrategia argumentativa [Irony as an argumentative strategy]. Revista Iberoamericana de Argumentación, 8, 1–15. Olgoso, A. (1999). Cuentos de otro mundo [Tales of another world]. Valladolid: Caja España. Paavola, S. (2004). Abduction through grammar, critic, and methodeutic. Transactions of the Charles S. Peirce Society: A Quarterly Journal in American Philosophy, 40(2), 245–270. Peirce, C. S. (1931–1958). In C. Hartshorne, P. Weiss, & A. W. Burks (Eds.), Collected papers, vols. 1–8. Cambridge, MA: Harvard University Press. (CP). Raccah, P.-Y. (2014). Linguistic argumentation as a shortcut for the empirical study of argumentative strategies. In B. Garssen, D. Godden, G. Mitchell, & F. Snoeck Henkemans (Eds.), Proceedings of 8th Conference of The International Society for the Study of Argumentation. Online publication: http://rozenbergquarterly.com/issa-proceedings-2014-linguisticargumentation-as-a-shortcut-for-the-empirical-study-of-argumentative-strategies/. Saul, J. M. (2002). What is said and psychological reality: Grice’s project and relevance theorists’ criticisms. Linguistics and Philosophy, 25, 347–372. Sperber, D., & Wilson, D. (1981). Irony and the use-mention distinction. In P. Cole (Ed.), Radical pragmatics (pp. 295–318). Nueva York: Academic Press. Tindale, C. W. (2015). The philosophy of argument and audience reception. New York: Cambridge University Press. van Eemeren, F. H., & Grootendorst, R. (1984). Speech acts in communicative discussions. Dordrecht, Netherlands: Foris. van Eemeren, F. H., & Grootendorst, R. (1987). Fallacies in pragma-dialectical perspective. Argumentation, 1, 283–301. van Eemeren, F. H., & Houtlosser, P. (2000). Rhetorical analysis within a pragma-dialectical framework. Argumentation, 14, 293–305. van Eemeren, F. H., & Houtlosser, P. (2009). Strategic maneuvering: Examining argumentation in context. In F. van Eemeren (Ed.), Examining argumentation in context (pp. 1–24). Amsterdam: John Benjamins B.V. West, D. E. (2016). The abductive character of Peirce’s virtual habit. Semiotics, Semiotics 2016: Archaeology of Concepts, 13–22. https://doi.org/10.5840/cpsem20165. Wilson, D. (2006). The pragmatics of verbal irony: Echo or pretence? Lingua, 116, 1722–1743. Wilson, D., & Sperber, D. (2004). Relevance Theory. In L. R. Horn & D. Ward (Eds.), The handbook of pragmatics (pp. 607–632). Oxford: Blackwell.

Chapter 14

Is “Conductive Argument” a Single Argument? Isabela Fairclough

14.1 Introduction This chapter is intended as a theoretical contribution to the literature on so-called “conductive argumentation”. I will focus only on those “conductive arguments” with a practical/normative conclusion (“Agent ought to do A”/“Proposal A is a reasonable course of action”), i.e. Wellman’s “third pattern”: “that form of argument in which some conclusion is drawn from both positive and negative considerations”, or in which “reasons against the conclusion are included as well as reasons for it” (Wellman, 1971, p. 57). In the argumentation literature (Blair & Johnson, 2011), this pattern is also called a “pro/con”, or a “balance-of-considerations” argument. I adopt a critical rationalist, anti-justificationist perspective on argument and decision-making (Popper, 1963, 1979; Miller, 1994, 2006, 2013, 2014). According to the philosophy of critical rationalism, arguments are “always negative; they are always critical arguments, used only and needed only to unseat conjectures that have been earlier surmised” (Miller, 1983, p. 10). In my view, therefore, what (confusingly) appears to be a “conductive argument”, as a distinct structure or (single) argument type, is a particular outcome of a process of critical testing of a proposal (a conjecture) with a view to making a decision in a rational way. Instead of seeing the proposal as finally supported by a set of “pro” reasons which have allegedly outweighed the “con” reasons, I suggest a view in which the proposal (conjecture) has successfully withstood the criticism directed against it. To say that the proposal has withstood criticism is to say that no compelling reasons why it should be abandoned have come to light. Whatever “cons” there are do not conclusively refute it, but whatever “pros” there are do not conclusively support it either. In light of its ability to achieve a desirable goal, the proposal has emerged as reasonable, having withstood criticism, and can be adopted (subject to conclusive objections coming I. Fairclough (B) University of Central Lancashire, Preston, UK e-mail: [email protected] © Springer Nature Switzerland AG 2020 F. H. van Eemeren and B. Garssen (eds.), From Argument Schemes to Argumentative Relations in the Wild, Argumentation Library 35, https://doi.org/10.1007/978-3-030-28367-4_14

223

224

I. Fairclough

to light at a later date, and subject to the discovery of a better alternative that has also withstood criticism). If rational decision-making involves a thorough search for potential reasons against a proposal, i.e. allowing for the possibility that some “con” reasons will conclusively support the opposite conclusion, it follows that two conclusions (not just one) are always potentially in play. Following Miller (1994, 2006, 2013), I take “rational decision-making” to mean the “rational making of decisions”, i.e. according to a methodical procedure of critical testing, and not the “making of rational decisions”. From a critical rationalist perspective, the conclusion of a so-called “conductive” argument is a conjecture (an informed “guess”) that a particular action is the right thing to do, in light of an intended goal, based on all current knowledge. This “conclusion” is not supported by an argument from goals or from positive consequence except in a defeasible way: it does not follow that one should do A only because A is capable of achieving some desirable goal or positive consequence. If, however, based on all current knowledge, the proposal (A) is not likely to achieve the goal (G), it should be abandoned. But, assuming that (based on all current knowledge) the proposal will achieve the goal (its intended consequence), its unintended consequences or side effects may still conclusively refute it, in case they are unacceptable and thus ought to be avoided. The proposal will withstand criticism, and emerge as a potentially reasonable course of action, if no unacceptable (intended or unintended) consequences have come to light while considering whether to adopt A or not. The form of argumentation involved when the conjecture that A is the right thing to do is refuted by its potential unacceptable consequences is therefore deductively valid—modus tollens: Action A will lead to consequence C. Non-C should be the case. Therefore, non-A should be the case. Conductive argumentation of the sort I am exploring here occurs in deliberative activity types (both deliberation with oneself or with others), and instantiates the abstract genre of deliberation.1 Deliberating agents put forward (alternative) proposals for action, conjecturing that these might help them achieve their goals. For decision-making to be rational, the agents should subject these proposals to criticism. The decision to adopt a particular proposal A will be reasonable if the hypothesis that A is the right course of action has been subjected to critical testing in light of all the knowledge available (including the perspectives of all those potentially affected) and has survived criticism (Fairclough & Fairclough, 2012; Fairclough, 2016, 2019). “Consequences” include known impacts, but also risks (which may not materialize). In addition, the situation where a proposal would clash with, or go against a moral or institutional principle, rule or norm, would also be an unacceptable consequence, 1 Van Eemeren (2010, pp. 138–143) distinguishes among genres, activity types and concrete speech

events. A particular debate (e.g. on a particular policy proposal), as speech event, instantiates the abstract category of policy debate as activity type, which in turn instantiates the abstract genre of deliberation. The intended outcome of deliberative activity types is a normative-practical conclusion (judgment) that can ground decision and action.

14 Is “Conductive Argument” a Single Argument?

225

on my account, should those principles, rules and norms be assessed (either singly or collectively) as ultimately non-overridable in a particular deliberative context.2 Any practical proposal A is likely to have a number of undesirable consequences. If, judging from the perspective of those who would be affected by the proposal, these undesirable consequences are not unacceptable, the deliberating agents might conclude that “Proposal A is reasonable, although—in addition to achieving some goals and resulting perhaps in other positive effects—it might result in some undesirable consequences”, i.e. in spite of counter-considerations. If, however, the potential consequences are deemed unacceptable, the agents may find it more reasonable to abandon the proposal. I will call reasons of this latter sort (e.g. unacceptable risks or impacts) “decisive objections”. I propose to distinguish therefore between two kinds of objections against a proposal: counter-considerations and decisive objections. This distinction is crucial to my account of “pro/con” argumentation. While testing various possible alternatives for achieving their goals, deliberating agents will consider reasons in favour (“pro”) and reasons against (“con”) each of these alternatives, in the form of the desirable and undesirable consequences they might result in. Not all the reasons against a proposal, as objections, will be assessed as having the same strength. Depending on how strong the objections are assessed to be, the agents may come to two main conclusions: (1) Proposal A is a reasonable course of action, despite the existence of a range of counter-considerations. The conclusion that A should not go ahead, always possible in principle, does not follow in this case, because none of the objections (“con” reasons) are deemed to be strong enough to warrant that conclusion: none are unacceptable (either singly or collectively). In the absence of decisive objections, and seeing as there are “pro” reasons (e.g. the proposal will achieve the desirable goal and possibly other positive consequences), the conclusion in favour of A can tentatively be maintained. (2) Proposal A is an unreasonable course of action because there are decisive objections against it (possibly in addition to counter-considerations), and in spite of a number of potentially positive consequences. In this case, the potential “conductive” argument tentatively supporting A (in light of its positive effects and despite counter-considerations) will collapse into a deductive argument in support of not doing A. The conclusion in favour of A can no longer be maintained (not even tentatively). The gist of my account is therefore the following: a practical proposal that has withstood critical testing (i.e. no decisive objections have emerged against it, capable of rebutting it), together with both the “pro” and “con” reasons that have emerged as relevant to it during deliberation (with oneself or with others), is indistinguishable from a so-called “conductive” argument in favour of that proposal. What appears to be a “conductive” argument is only a recapitulation or summary of the deliberative 2 The

critical rationalist view of deliberation and decision-making adopted here is different but not incompatible with the frameworks for deliberation proposed by Hitchcock, McBurney and Parsons (2001) and McBurney, McBurney, Hitchcock and Parsons (2007).

226

I. Fairclough

process, as it has unfolded in time, in a particular situation. This is the situation where, having examined both “pro” and “con” reasons, and having concluded, by asking and answering critical questions, that none of the “con” reasons are decisive objections but only counter-considerations, and also in light of proposal’s potential to achieve the stated goals and possibly other positive effects, the deliberating agents have concluded that proposal A can be maintained. What is meant by this is nothing more than: proposal A has so far survived criticism and can be provisionally accepted as a reasonable alternative, subject to decisive objections coming to light at a later date. Saying that the proposal has survived criticism means that no decisive objections against it have come to light. It does not mean that there are no reasons against it at all: there may indeed be several such reasons, but they have been assessed as mere counter-considerations, not decisive objections.

14.2 The Current Consensus: Representing “Conductive Argument” as a Single Argument Interest in the concept of “conductive” argument was revived by Govier (1999, 2001, 2011a, 2011b), who suggested that counter-considerations to a conclusion should be represented as “wavy lines” or “wavy lines and a bar” (Fig. 14.1). On her view, “counter-considerations should not be regarded as premises of an argument, because they do not support the conclusion and are not put forward by the arguer as supporting the conclusion”. She also suggests that the “weighing” metaphor adequately captures what goes on in conductive reasoning. Namely, a person who acknowledges counter-considerations, and nevertheless still wishes to put forward the argument that his conclusion is supported by positively relevant premises, is committed to the judgment that the supporting (positively relevant) premises outweigh the counterconsiderations. In other words, “although there are reasons for [the] conclusion and reasons against it, the reasons for it are stronger and more convincing than the reasons against it” (Govier, 2001, p. 396). In a more recent collection of proposals on how to define, analyze, represent and assess “conductive arguments” and “conductive reasoning”, published by Blair and Johnson (2011), Hansen (2011) has suggested that “conductive arguments” contain an implicit on-balance (OB) premise, saying that on balance, the pros outweigh the cons, thus acknowledging that the weighing of pro/con reasons is an essential part of such arguments (Fig. 14.2). Fig. 14.1 Representing conductive arguments (taken from Govier, 2001, p. 395)

14 Is “Conductive Argument” a Single Argument?

227

Fig. 14.2 Representing conductive arguments (taken from Hansen, 2011, p. 40)

Fig. 14.3 Representing conductive arguments (taken from Jin, 2011, p. 29)

In the same volume, integrating proposals for the representation of pro/con arguments made by Govier (1999), and of objections by Hitchcock (1983), Jin (2011) suggested the diagram in Fig. 14.3. Here, two considerations are independently supporting the conclusion, and two considerations are independently opposing it. Even if the latter are acknowledged by the arguer, the conclusion can still be maintained, because the cumulative support of the “pro” considerations outweighs the cumulative objection of the counter-considerations. The “pro” reasons manage to “shunt aside” (Hitchcock’s term) the “con” reasons. According to all these proposals, “conductive argument” is a single argument: a single conclusion is supported by a range of reasons, despite the existence of reasons against. By contrast, I am sugesting that so-called “conductive argument” is not a single argument, and that the opposite (negative) conclusion should also be represented, in relation to the positive one. As I show below, the situation where the positive conclusion can be maintained, in spite of the reasons against, is only one possible configuration or outcome of the deliberative process. It is always possible that the positive conclusion will not survive criticism. In this case, the argument in favour will collapse into a deductive argument in favour of the opposite conclusion, and the “conductive argument” will disintegrate under the weight of the decisive objections brought against its positive conclusion.

14.3 My Proposal: Representing Pro/Con Argumentation as Deliberative Process Unlike the view I defend here, all of the above proposals assume that “conductive argument” is a single argument. A single conclusion is defended, for reasons X and Y, despite the existence of reasons Z and W. This is because reasons X and Y (the “pros”) are said to outweigh Z and W (the “cons”). Because a “conductive” argument is taken to be a single argument, and the negatively relevant reasons cannot

228

I. Fairclough

by definition support the conclusion, but count against it, these reasons (generically called “counter-considerations”) are not (and could not be) “premises” (Govier, 2001, p. 396). What I suggest is the following: (1) There is no “conductive argument”, if what is meant by the term is a single argument, consisting of one conclusion and two sets of reasons (positively and negatively relevant to that conclusion). Instead, two possible conclusions (“A is reasonable” and “A is not reasonable”) are always potentially in play, and which one will survive criticism is to emerge at the end of a process of critical questioning. For practical/normative conclusions, critical questioning takes place in a deliberative generic framework (Fairclough & Fairclough, 2012), and the acceptability of the conclusion is tested in the light of its possible or known consequences, and of how it might survive in conditions of uncertainty. A critical rationalist perspective also acknowledges that a proposal with potentially unacceptable consequences need not be discarded if there is a satisfactory Plan B, insurance or mitigation strategy in place; moreover, such proposals may not be discarded because agents are willing to take the risk of unacceptable consequences materializing (Miller, 2013; Fairclough, 2015). (2) The “con” reasons are premises (what else could they be?) in an argument criticizing the proposal, and thus potentially supporting the opposite conclusion (“A is not reasonable”). If the “con” reasons, intended as objections to A, are found to be decisive objections, it will be more reasonable to abandon A. If these objections are found to be mere counter-considerations, it will not follow that A should be abandoned. Nor will it follow that it should be adopted, or that it is recommended, least of all that it is the best course of action, but merely (at this stage) that it is a reasonable course of action (among other possibly reasonable alternatives). The conclusion that A is the best among alternatives can only emerge from further critical questioning, once all unreasonable alternatives have been eliminated, and in light of some criterion providing a reasonable basis for choosing the “best” course of action among reasonable alternatives (Fairclough, 2015). (3) A crucial distinction must be drawn between counter-considerations, as objections that can be incorporated into the arguer’s case, and decisive objections, as objections that cannot be thus incorporated, but lead to the collapse of the case in favour of doing A. I have previously used the term “critical objections” for those objections that (unlike counter-considerations) can rebut the conclusion in favour of a proposal (Fairclough, 2015, 2018a, 2018b). To avoid confusion, given that all objections, weak or strong, are in a sense “critical”, I suggest changing the term “critical objection” to “decisive objection”, when referring to critical objections that rebut the conclusion.3 Thus, in a process of critical questioning, in dialogue, arguers may assess an objection aimed against a proposal as being a mere counter-consideration, in which case the case for A may 3I

am grateful to Erik Krabbe for pointing out the need to find a better name for ‘critical objection’ and to David Miller for suggesting ‘decisive objection’.

14 Is “Conductive Argument” a Single Argument?

229

still stand, or as a decisive objection, in which case the case for A is rebutted. In practical reasoning, potential consequences of a course of action that are deemed to be unacceptable will be decisive objections against it (though this may not be a sufficient reason for the agents to refrain from that course of action). The distinction between counter-considerations and decisive objections is emphatically not one and the same with that between counter-considerations and objections, which has been usefully clarified by Johnson (2011) and acknowledged as correct by Govier (2011b). If, originally for Govier (2001, p. 395), “counterconsiderations” and “objections” were one and the same thing, it has become clear in the meantime that it was in fact a mistake to conflate the two. Counter-considerations are claims negatively relevant (or taken to be negatively relevant) to the acceptability of the conclusion and acknowledged by the arguer to have that status. As such, counter-considerations are part of the arguer’s case. If, after being raised in a dialectical context, objections come to be acknowledged by an arguer and incorporated into an adapted argument, at that point they would play the role of counter-considerations (Govier, 2011b, p. 2).

This is also the view taken here. Objections are raised by a critic (an external critic or, in the case of monological deliberation, the arguer him/herself, as antagonist of his/her own standpoint), and may be incorporated into the arguer’s case as counterconsiderations (in which case the original standpoint can be maintained). Objections can also lead to the abandonment of the original standpoint, and the disintegration of the pro/con argument, in those cases when they cannot be incorporated into the original argument, because they are decisive objections that simply rebut the case in favour. My proposal is therefore to distinguish not only between objections (as raised by a critic) and counter-considerations (as assessed by the arguer)—as Johnson (2011) has done—but also between objections, as raised by a critic (as antagonist), and two different ways the critic’s objections may be assessed by the arguer (as protagonist): (a) as counter-considerations, and (b) as decisive objections. If proposal A has survived criticism in light of the critic’s objections, then those objections can be incorporated into the final argument as counter-considerations. If the objections have been deemed to have a much stronger critical force, to be decisive objections (e.g. the consequences of A, if performed, would be unacceptable), then those objections, as rebuttals, will conclusively support the opposite conclusion, and A should be abandoned. These two distinctions are captured in Fig. 14.4. To sum up, in deliberative activity types, in a context of dialogue, objections may be raised by a critic against a proposal that is being tested. The critic may point out, for example, that A is likely to have an undesirable, possibly even an unacceptable consequence. Assuming the arguer acknowledges this objection, including the particular way it is formulated (or ‘framed’),4 as relevant to his case, he can assess its critical force in two main ways: 4 Arguers (or the deliberating agents involved in the critical testing of a proposal) may not acknowl-

edge their critics’ objections, including the particular way they are framed. I focus here on the simple case in which the same objections are acknowledged by both parties, and there is no incompatible

230

I. Fairclough

Fig. 14.4 The antagonist’s objections (left) are evaluated by the protagonist (right) as either counter-considerations or as decisive objections

(1) as a (mere) counter-consideration, incapable of rebutting the conclusion, in which case he may still maintain his commitment to the proposal and can incorporate the objection into his original case (“proposal A still stands in spite of these counter-considerations”). (2) as a decisive objection (or rebuttal), in which case he will no longer be able to maintain his commitment to the proposal; a decisive objection cannot be incorporated into the original case in favour of the conclusion, but indicates that conclusion is false, and the proposal must be abandoned. A decisive objection conclusively supports the opposite, negative conclusion (“proposal A is not reasonable”). The assessment of an objection as being a counter-consideration or a decisive objection is the result of a process of critical questioning. Once an objection is raised, the arguer may or may not see immediately that it is decisive and rebuts his case, even if later on, as a result of further dialogue, he might assess it as such. Alternatively, several objections that are individually assessed as mere counter-considerations may end up being deemed to rebut the positive conclusion through their cumulative force. In this case, it may not be reasonable to go ahead with A in the face of so many counterconsiderations, though none was individually assessed as a decisive objection. What objections there are against a proposal and how strong they are is not known entirely in advance, but will emerge in a process of critical questioning, unfolding in time, divergence between the way these objections are framed on each side. In the case discussed here, there can be intersubjective disagreement (resolvable or not) concerning the strength of objections (are they mere counter-considerations or decisive objections?). I do not discuss the more complex case where there is divergence in the way the pro/con reasons are framed (discussed by Wohlrapp 1998, 2011), nor the case where the pro/con reasons advanced by one side are not recognized as relevant by the opposite side.

14 Is “Conductive Argument” a Single Argument?

231

within a deliberative process, in light of all the information available, and taking into account the perspectives of all those who would be affected by the proposal. On a critical rationalist perspective, practical statements (like theoretical ones) can be provisionally accepted or conclusively rejected based on how they withstand criticism in light of their potential consequences. Let us assume that one alternative A1 is being tested among several possible (A1 …An ). Figure 14.5 represents both potential conclusions (A1 and non-A1 ) and which one will be actualized depends on how they will fare in the process of critical testing. Clearly, both conclusions cannot both be true at the same time, and one will be discarded at the end of the deliberative process. The conclusion on the left side (Proposal A1 is not reasonable) will follow conclusively if it is the case that one or more objections have been (singly or collectively) assessed as decisive objections. The conclusion on the right (Proposal A1 is reasonable) can be tentatively, provisionally maintained if it survives criticism, that is, if it is indeed the case that no objections, singly or collectively, are decisive objections, but merely counter-considerations. A proposal can only be assessed as potentially reasonable if, having gone through a process of critical testing to the best of the deliberating agents’ knowledge, it has withstood criticism (i.e. no decisive objections have emerged against it). In the diagrams below, the statements in the boxes shown in bold are added, as premises, to the original premise set as a result of the process of critical questioning, as it unfolds during deliberation. Figure 14.6 represents the situation where the case in favour of doing A is rebutted by one or more decisive objections that have come to light by critical questioning. The argumentative configuration previously shown in Fig. 14.5, where either conclusion was in principle possible, before critical questioning was allowed to proceed, has now collapsed into a deductive argument in favour of non-A (left-hand side). Namely, if it were true that the potential consequences (singly or collectively) were unacceptable for one or more of the parties that would be affected, then it would follow conclusively that A is not a reasonable course of action. While the statements on the right-hand side (the positive consequences) may still be true, no argument (not even a defeasible one) can be made starting from these statements as premises (and the blank box where the “pro” conclusion would have been is supposed to suggest this). The positive conclusion (Proposal A1 is reasonable, which is possible in principle) cannot be drawn at all any more, not even defeasibly. In the absence of a premise saying that there are no decisive objections, it can no longer be maintained that the proposal is reasonable. That conclusion is in fact rebutted by the fact that there are such objections. By contrast, Fig. 14.7 captures the situation when, at the end of a process of critical questioning, to the best of the critics’ knowledge, no decisive objections have come to light. The proposal has survived criticism: the objections have been assessed as (mere) counter-considerations. In this case, it does not follow that the proposal should be abandoned. The proposal can be provisionally, tentatively, adopted, subject of course to stronger, decisive objections coming to light at a later date. The conclusion that A is not reasonable (left), while possible, in principle, has not obtained in this case

232 Fig. 14.5 Deliberation scheme: Proposal may be rebutted in light of decisive objections (left) or may survive criticism if there are no decisive objections (right)

I. Fairclough

14 Is “Conductive Argument” a Single Argument? Fig. 14.6 There are decisive objections to the proposal: it follows conclusively that it should be abandoned

233

234 Fig. 14.7 There are no decisive objections, only counter-considerations; proposal A can be tentatively maintained

I. Fairclough

14 Is “Conductive Argument” a Single Argument?

235

(again, the blank box where this negative conclusion would have been is supposed to represent this). In both Figs. 14.6 and 14.7, the statements that used to be premises in the arguments shown in Fig. 14.5, but no longer “support” any conclusion and are left stranded, can be included in a summary of the deliberating process. Having assessed both “pro” and “con” reasons, deliberating agents may conclude that: (1) although it has potential desirable consequences (“pro” reasons), proposal A is ultimately unreasonable and ought to be rejected, because (among the “con” reasons) there are decisive objections against it (Fig. 14.6)—i.e. the proposal has not survived criticism, or: (2) although it has potential undesirable consequences (“con” reasons), proposal A is reasonable and can be provisionally maintained, because (among the “con” reasons) there are no decisive objections against it (Fig. 14.7)—i.e. the proposal has survived criticism, and there is no compelling reason, at least for the time being, why it should be abandoned. It is easy to see how the latter summary captures the logic of a so-called “conductive argument” in favour of proposal A. This summary, which reduces a process to its a-temporal end-result, is in fact virtually indistinguishable from what is commonly called “conductive argument”. However, to speak of this process—unfolding over time and involving the raising and assessment of objections in a context of dialogue—as a “single argument” (and somewhat on a par with deductive and inductive arguments, as a distinct “type” of argument) is in my view a category mistake.

14.4 Conclusion Deliberation occurs in time. Arguers put forward proposals in view of goals, values and the situation they are in. Proposals are not strictly speaking supported or justified by these premises, but are informed guesses, hypotheses, or conjectures that some course of action might solve the problems and achieve the goals. Deliberating with themselves or with others, rational arguers will subject these conjectures to critical testing. What consequences might come about, assuming they went ahead with a proposed course of action? Would these consequences be ultimately acceptable? The answers provisionally given to these questions may themselves be questionable. There may be disagreement between the arguers and their critics, or between deliberating agents, on the severity of a particular consequence. What is a merely an undesirable but ultimately an acceptable consequence to someone may be totally unacceptable to someone else. At the end of a process of critical questioning, arguers will either conclude that, at least for them, and to the best of their knowledge, a particular proposal that seems capable of achieving their goals may be maintained, given that there are no decisive objections against it, or they may conclude that it has to be discarded, if there are such objections.

236

I. Fairclough

I have argued against seeing “conductive argument” as a single argument, and suggested a view of deliberative dialogue, involving a process of critical testing of proposals, in which objections are raised and their critical force is assessed in light of all the knowledge at the deliberators’ disposal, and from the perspectives of all those affected. A possible deliberative outcome is the conclusion that a particular proposal must be discarded as unreasonable, when the objections that have been raised against it are evaluated as decisive objections, not mere counter-considerations. This is to say that the consequences of that proposal would be ultimately unacceptable. Having decided to abandon a proposal, deliberating agents can start testing another alternative. But whenever a proposal has survived criticism, i.e. whenever it is not rebutted by decisive objections, and whatever objections have been raised were only assessed as counter-considerations, then a recapitulation of the whole process that deliberating agents have engaged in, including the reasons in favour and the reasons against the proposal, will take the form of what has been conventionally called a “conductive” argument. This (instant) summary, at the end of a (temporal) process whose purpose has been the critical testing of a proposal, is in fact one and the same with a so-called “conductive” argument in favour of that proposal.

References Blair, J. A., & Johnson, R. H. (Eds.). (2011). Conductive argument, an overlooked type of defeasible reasoning. London: College Publications. Fairclough, I. (2015). A dialectical profile for the evaluation of practical arguments. In B. Garssen, D. Godden, G. Mitchell & A. F. Snoeck Henkemans (Eds.), Proceedings of the 8th International Conference of the International Society for the Study of Argumentation. Amsterdam: SicSat. Fairclough, I. (2016). Evaluating policy as practical argument: The public debate over the first UK austerity budget. Critical Discourse Studies, 13(1), 57–77. https://doi.org/10.1080/17405904. 2015.1074595. Fairclough, I. (2018a). Deliberative discourse. In J. Richardson & J. Flowerdew (Eds.), The Routledge handbook of critical discourse analysis (pp. 242–256). London: Routledge. Fairclough, I. (2018b). Conductive argumentation in the UK fracking debate. In S. Oswald & D. Maillat (Eds.), Argumentation and inference: Proceedings of the 2nd European Conference on Argumentation, Fribourg 2017, Vol. II (pp. 297–310). London: College Publications. Fairclough, I. (2019). Deontic power and institutional contexts: The impact of institutional design on deliberation and decision-making in the UK fracking debate. Journal of Argumentation in Context, 8(1), 136–171. https://doi.org/10.1075/jaic.18014.fai. Fairclough, I., & Fairclough, N. (2012). Political discourse analysis. London: Routledge. Govier, T. (1999). The philosophy of argument. Newport News, VA: Vale Press. Govier, T. (2001). A practical study of argument (5th ed.). Belmont, CA: Wadsworth. Govier, T. (2011a). More on counter-considerations. In F. Zenker (Ed.), Argumentation: Cognition and community. Proceedings of the 9th International Conference of the Ontario Society for the Study of Argumentation (OSSA). CD ROM (pp. 1–10). University of Windsor. Govier, T. (2011b). Conductive arguments: Overview of the symposium. In J. A. Blair & R. H. Johnson (Eds.), Conductive arguments, an overlooked type of defeasible reasoning (pp. 262–276). London: College Publications.

14 Is “Conductive Argument” a Single Argument?

237

Hansen, H. V. (2011). Notes on balance-of-consideration arguments. In J. A. Blair & R. H. Johnson (Eds.), Conductive arguments, an overlooked type of defeasible reasoning (pp. 31–51). London: College Publications. Hitchcock, D. (1983). Critical thinking: A guide to evaluating information. Toronto: Methuen. Hitchcock, D., McBurney, P., & Parsons, S. (2001). A framework for deliberation dialogues. In H. V. Hansen, C. W Tindale, J. A. Blair, & R. H. Johnson (Eds.), Argument and Its Applications: Proceedings of the Fourth Biennial Conference of the Ontario Society for the Study of Argumentation (OSSA 2001). Jin, R. (2011). The structure of pro and con arguments: A survey of the theories. In J. A. Blair & R. H. Johnson (Eds.), Conductive arguments, an overlooked type of defeasible reasoning (pp. 10–30). London: College Publications. Johnson, R. H. (2011). The relationship between pro/con and dialectical tier arguments. In J. A. Blair & R. H. Johnson (Eds.), Conductive arguments, an overlooked type of defeasible reasoning (pp. 52–61). London: College Publications. McBurney, P., Hitchcock, D., & Parsons, S. (2007). The eightfold way of deliberation dialogue. International Journal of Intelligent Systems, 22(1), 95–132. Miller, D. W. (Ed.). (1983). A pocket popper. Fontana Paperbacks. Oxford: Oxford University Press. Miller, D. W. (1994). Critical rationalism: A restatement and defence. Chicago: Open Court. Miller, D. W. (2006). Out of error. Further essays on critical rationalism. London: Routledge. Miller, D. W. (2013). Deduktivistische Entscheidungsfindung. In R. Neck & H. Stelzer (Eds.), Kritischer rationalismus heute. Zur aktuaklität de philosophie Karl Poppers, Schriftenreihe der Karl Popper Foundation (pp. 45–78). Frankfurt am Main: Peter Lang. German translation by Christina Kopetzky of Deductivist Decision Making (unpublished MS). Miller, D. W. (2014, Enero–Junio). Some hard questions for critical rationalism. Discusiones filosoficas (Manizales), 15, 24, 15–40. Popper, K. R. (1963). Conjectures and Refutations (5th ed.). London: Routledge and Kegan Paul, 1989. Popper, K. R. (1979). Objective knowledge. An evolutionary approach. Oxford: Oxford University Press. van Eemeren, F. H. (2010). Strategic maneuvering in argumentative discourse. Amsterdam: John Benjamins. Wellman, C. (1971). Challenge and response: Justification in ethics. Carbondale, IL: Southern Illinois University Press. Wohlrapp, H. (1998). A new light on non-deductive argumentation schemes. Argumentation, 12, 341–350. Wohlrapp, H. (2011). A misleading model for the analysis of pro- and contra-argumentation. In J. A. Blair & R. H. Johnson (Eds.), Conductive arguments, an overlooked type of defeasible reasoning (pp. 210–223). London: College Publications.

Chapter 15

On the Logical Reconstruction of Conductive Arguments Yun Xie

15.1 Introduction In a book published in 1971, Challenge and Response: Justification in Ethics, Carl Wellman introduces the notion of “conduction” to refer to a particular type of reasoning. As a moral philosopher, Wellman claims that there are certain arguments in ethics that “are left over when all deductive and inductive ethical arguments have been studied” (Wellman, 1971, p. 51). Those arguments, according to him, represent a type of reasoning that “derives its conclusion from a variety of premises each of which has some independent relevance” (p. 51). Hence, he contends that “since what is characteristic of this sort of reasoning is the leading together of various considerations, it seems appropriate to label it ‘conduction’.” (p. 51) In order to further clarify the nature of conduction, Wellman defines it as a sort of reasoning “in which (1) a conclusion about some individual case (2) is drawn non-conclusively (3) from one or more premises about the same case (4) without any appeal to other cases.” (p. 52) On that basis, three patterns of conduction are distinguished (pp. 55–57), but only the third one has received much attention in argumentation literature, thus becomes the representative form of conductive arguments. The third pattern of conduction is that form of argument in which some conclusion is drawn from both positive and negative considerations. In this pattern reasons against the conclusion are included as well as reasons for it. For example ‘in spite of a certain dissonance, that piece of music is beautiful because of its dynamic quality and its final resolution’ or ‘although your lawn needs cutting, you ought to take your son to the movies because the picture is ideal for children and will be gone by tomorrow. (Wellman, 1971, p. 57)

The distinctive feature of such a pattern of conduction, as is indicated here by Wellman, is the particular way of reaching a conclusion from both reasons for and Y. Xie (B) Department of Philosophy, Institute of Logic and Cognition, Sun Yat-sen University, Guangzhou, China e-mail: [email protected] © Springer Nature Switzerland AG 2020 F. H. van Eemeren and B. Garssen (eds.), From Argument Schemes to Argumentative Relations in the Wild, Argumentation Library 35, https://doi.org/10.1007/978-3-030-28367-4_15

239

240

Y. Xie

against it. In fact, this pattern does capture a special way of arguing that is commonly recognized in practice, especially in the contexts where we are arguing about issues that have a pro-con nature. Wellman truly believes that conduction represents a type of argument that is neither deductive nor inductive (Wellman, 1971, p. 51). This idea is well endorsed few years later by many argumentation scholars, and they have further contended that conductive arguments should be taken as “an overlooked type of defeasible reasoning” (Blair & Johnson, 2011; Govier, 1979, 1987). In general, the notion of conductive argument is of particular value in argumentation studies, for there seems to be no other type of argument that could explicitly collect both affirmative and negative reasons bearing on the conclusion into a single structure. It is then a new and challenging task to explain why we collected both positive and negative considerations into one structure, and how it works to establish a standpoint and to achieve persuasion. In the last decade, there is a variety of theoretical accounts on conduction that are developed out from different perspectives of logic (Blair, 2016; Freeman, 2011; Govier, 2010; Hansen, 2011), dialectics (Juthe, 2018; van Laar, 2014) and rhetoric (Xie, 2017). This paper is a reflection about the adequacy of the logical account on conduction, focusing particularly on the issue of the logical structure of conductive arguments. Specifically, it aims to critically discuss the two main approaches for the logical reconstruction of conductive arguments that are currently prevailing in the field. The first one is the supplementation of on-balance premise approach, as developed in Hansen (2011) and Blair (2017), and the second is the warrant-reformulation approach, as proposed by Freeman (2011) and Bermejo-Luque (2017). It is argued that neither of them is satisfactory, at least in their versions developed till now. In Sect. 15.2, I first explore these two approaches in more detail, and then in Sect. 15.3 I offer a criticism to the warrant-reformulation approach by examining the shortcomings in the accounts developed respectively by Freeman and Bermejo-Luque. After that, in Sect. 15.4 and 15.5, I discuss two main problems in the supplementation of on-balance premise approach, and conclude in Sect. 15.6.

15.2 The Logical Reconstruction of Conductive Arguments 15.2.1 The Supplementation of On-balance Premise Approach For most scholars who have taken a logical perspective on conduction, their views are strongly influenced by Wellman’s work. In his discussions on conductive arguments, Wellman claims that the third pattern of conduction “raises the question of how one knows that the reasons for the conclusion are stronger than those against it” (Wellman, 1971, p. 57). In this connection, he further indicates that “perhaps the most popular model for this sort of conductive reasoning is weighing. One decides whether the argument is valid by weighing the pros against the cons.” (p. 57) And

15 On the Logical Reconstruction of Conductive Arguments

241

as his final verdict, Wellman contends that “on the whole it is helpful to think of conductive reasoning in terms of the model of weighing” (p. 58). It is easy to see that here Wellman has brought about the idea of conceiving conductive arguments in terms of a mechanism of weighing. Particularly, the inclusion of the reasons against the conclusion is regarded as an indicator for a special means of justification, i.e., the conclusion is justified by weighing between the reasons for and against it. This characterization of conduction is widely accepted by many informal logicians, and becomes their starting point for understanding the logical structure of conductive arguments. For them, although in its appearance a conductive argument only includes some counter-considerations as its constituent part, there is always a need to formulate another implicit “on-balance premise” for it. An on-balance premise will clearly state that the reasons for the conclusion have outweighed those counter-considerations in justifying the conclusion. Therefore, the adding of such an unstated premise to a conductive argument not only makes its underlying process of weighing explicit, but also completes the process with a key component. Moreover, such an addition is also plausible because when putting forward a conductive argument “an arguer is committing himself or herself to the claim that the supporting considerations ‘outweigh’ the counter-considerations and render the conclusion acceptable” (Govier, 2011, p. 266, italics original). In line with these considerations, the supplementation of on-balance premise has then been taken as a proper way for reconstructing conductive arguments. In a recent paper, Blair has proposed a model of conductive arguments in the following form (2016, p. 124): 1.1 a, b, c,…, support p. 1.2 w, x, y,…, support not-p. 1.3 a, b, c,… outweigh w, x, y,… (or conversely). So, 1 p (or not-p).

This can be taken as the typical version of the supplementation of on-balance premise approach to the reconstruction of conductive arguments . Obviously, in this model the premise 1.3 is the on-balance premise added to the argument, and it states specifically an outweighing relation between reasons for and against the conclusion. In a similar vein, Hansen (2011) offers another more complicate model for conductive arguments that includes two sub-arguments (p. 39): P1 : Independent reason1 (for conclusion K) …… Pn : Independent reasonn (for conclusion K) Pn+1 : The reasons in P1 to Pn taken together outweigh the independent counterconsiderations to K, CC 1 to CC n taken together Conclusion: K even though CC 1 & …& CC n (inference to ‘even though’) Premise: K even though CC 1 & …& CC n Conclusion: K (simplification)

Here it is easy to see that the premise Pn+1 is exactly the on-balance premise supplemented in reconstruction. As Hansen has suggested, all conductive arguments

242

Y. Xie

“may be viewed as having such an implicit on-balance premise” (p. 39), for it needs to be presumed for us to make an inference to the claim p even though q, and both the on-balance premise and the claim p even though q should be taken as necessary parts of the reasoning in a conductive argument.

15.2.2 The Warrant-Reformulation Approach There are also some other scholars who have developed their views on conduction without endorsing Wellman’s model of weighing. Among them the most notable account is proposed by Freeman (2011) in light of the Toulmin Model. As is noted by Freeman, “Wellman’s insights can be preserved and better appreciated through both a generalization of his notion of conductive argument and through asking how such arguments might be ‘laid out’ on the Toulmin model”(p. 127). Since the Toulmin model is a field-invariant tool that characterizes the layout of any argument across all the disciplines, it also provides a paradigm for conceiving the structure of conductive arguments. And given that a conductive argument differs from other arguments only in its inclusion of some counter-considerations, then the key issue in reconstructing it by Toulmin model is to specify the particular function of counter-considerations, and thereby to correctly identify them as an element in the model. In regard to the question of “where do counter-considerations fit into the Toulmin model”, Freeman contends that “we can understand the function of these counterconsiderations more accurately and represent them more fully if we regard them as rebuttal.” (Freeman, 2011, p. 139) However, when reconstructing a conductive argument, he then comes out with a proposal that is slightly different. According to him, since in a conductive argument “the proponent did not reply to these counter-considerations, but simply conceded them. Hence we include the counterconsiderations as conjunctions in the premise of the warrant” (p. 140). To illustrate this idea, the following argument is given as an example: (1) I travel 350 miles each week by car. Even though (2) in summer I pick fresh vegetables and fruits from my own garden, (3) in winter I buy them from health food stores that truck them east from organic farms in California. Again, even though (4) I write on both sides of the paper, (5) I use paper and a great deal of it. Hence, (6) I use waste making energy. (Freeman, 2011, p. 138)

This conductive argument, in Freeman’s analysis, has a warrant as the following: From: x travels 350 miles each week by car & x in summer picks fresh fruits and vegetables from x’s garden & in winter x buys fruits and vegetables trucked east from California & x writes on both sides of the paper & x uses a great deal of paper To infer ceteris paribus: x uses waste making energy (p.140)

In fact, in this reconstruction we could find that counter-considerations in the conductive argument are first recognized, together with all the positive reasons, as the data of the argument. If compared with a non-conductive argument, the presence

15 On the Logical Reconstruction of Conductive Arguments

243

of counter-considerations enriches the data of the argument, therefore makes its inference rule become more complicate. As a result, all the counter-considerations should be conjoined in the premise of the warrant, in order to properly sanction the inference from the data to the claim. In a recent paper, Bermejo-Luque (2017) has also proposed a Toulminian account with regard to the reconstruction of conductive arguments. Based on a particular understanding on the nature of warrant, Bermejo-Luque has combined the Toulmin Model with the speech act theory, and thereby developed it into a linguistic normative model of argumentation (Bermejo-Luque, 2011). Using the theoretical framework pertaining to that model, she provides an account on the structure of conduction that is quite different from Freeman’s. According to Bermejo-Luque, the use of a conductive argument like “Although your lawn needs cutting, you ought to take your son to the movie because the picture is ideal for children and will be gone by tomorrow” should be analyzed as consisting of “two speech-acts: the concessive speech-act ‘your lawn needs cutting’, and the speech-act of arguing ‘you ought to take your son to the movies because the picture is ideal for children and will be gone by tomorrow’” (Bermejo-Luque, 2017, p. 11). Accordingly, such a conductive argument will be reconstructed in the following way: Premise (representing the act of adducing): (it is true that) the picture is ideal for children and (it is true that it) will be gone by tomorrow Conclusion (representing the act of concluding): (plausibly) (it is true that) you should take your son to the movie Warrant (representing the inference-claim): (it is plausible that) if it is true that the picture is ideal for children and it is true that it will be gone by tomorrow, then it is true that you should take your son to the movie (Bermejo-Luque, 2017, p. 12)

Here it is easy to see that in this reconstruction all the counter-considerations have just been removed. As is explained by Bermejo-Luque, since the concessive speech-act is of a different kind and different order from that of arguing, it could not be taken as conveying something that is part of the argument, therefore, “counter-considerations remain as an independent part of the argument, for they would play no properly argumentative role” (Bermejo-Luque, 2017, p. 13). However, as she has further clarified, counter-considerations are assigned an active role in evaluating the conductive argument, for they are additional important information that is relevant for us to determine the plausibility of warrant in the conductive argument (Bermejo-Luque, 2017, p. 12).

15.3 The Linguistic Feature of Conductive Arguments It is obvious that both Freeman and Bermejo-Luque have taken counterconsiderations in a conductive argument as something related to the notion of warrant in the Toulmin model, but they differ in understanding their status within the argument itself. Freeman stresses that counter-considerations are conceded by the

244

Y. Xie

arguer, hence he considers them as an integral part of the argument, and then also as a constituent element of the warrant. But Bermejo-Luque believes that the counterconsiderations, though conceded, have no bearing on the speech-act of arguing, therefore, she takes them as a separate unit that only has some value in assessing the plausibility of warrant. Moreover, “counter-considerations are not part of the warrant, even though the plausibility of this conditional depends on whether or not, overall, counter-considerations do not overwhelm pro-considerations”(Bermejo-Luque, 2017, p. 13). As Bermejo-Luque has explained, a counter-consideration should be regarded as some additional information that is relevant to the plausibility of the warrant, “even if no one had pointed at it”, and it is “something that no assessment of argumentation can escape” (Bermejo-Luque, 2017, p. 12). In other words, for Bermejo-Luque, counter-considerations are just necessary background information that is needed in the evaluation of any argument, their presence in a conductive argument makes no difference in its argument structure. Nor does it matter which counter-considerations are included there in a conductive argument (see BermejoLuque’s examples in 2017, p. 14). I think that both Freeman and Bermejo-Luque have failed in their accounts to give full credit to the language uses of the arguer who has advanced a conductive argument, and as a result their reconstructions simply make conduction trivialized as an argument type, and indistinguishable from many other arguments. As noted in Sect. 15.1, since the inception of the notion, conduction is characterized as a pattern in which reasons for and against the conclusion are together provided in an argument. Specifically, the inclusion of counter-considerations in a conductive argument is realized in particular by the use of an “even though”, “although”, or “notwithstanding” clause, and this linguistic feature has also been taken as a unique character of conductive arguments. It goes without saying that the introducing of counter-considerations in such a special way is definitely part of the arguer’s intent, because there are indeed several different possible ways to mention a counter-consideration in argument but the arguer has made such a choice among them. From a pragmatic point of view, the arguer’s introducing counter-considerations by those even-though type clauses serves to convey the conventional implications that “these reasons against the conclusion are outweighed” (Adler, 2013, p. 247), which, in turn, reveals explicitly that for the arguer the counter-considerations are not just conceded, but has also been taken as overridden and downplayed in importance. In light of these observations, it can be seen that Freeman’s account rightly admits the arguer’s concession of counter-considerations when presenting them in a conductive argument, but overlooks the fact that the arguer has also regarded them as outweighed. Hence in his reconstruction such an intended imbalance between pro and con considerations would simply disappear. As we have seen from his formulation of the warrant for conductive arguments, all the counter-considerations are just taken as equal conjuncts with all the reasons for the conclusion. Therefore, such a reconstruction of conductive arguments would turn out to be based upon a singlefaceted interpretation of their linguistic feature. Accordingly, it will just eliminate the distinctiveness of conduction pertaining to the indicators like “even though”, and thereby makes it indistinguishable from those arguments which acknowledge or

15 On the Logical Reconstruction of Conductive Arguments

245

concede counter-considerations by the means other than “even though” clause. To illustrate this point, we could take Freeman’s example above, and revise it into the following one by deleting all the uses of “even though”: (1) I travel 350 miles each week by car, (2) in summer I pick fresh vegetables and fruits from my own garden, (3) and in winter I buy them from health food stores that truck them east from organic farms in California. Again, (4) I write on both sides of the paper, and (5) I use paper and a great deal of it. So, anyway, to sum up, (6) I use waste making energy.

In line with Freeman’s reconstruction, this argument would have exactly the same warrant with the original conductive argument that has “even though” clauses: From: x travels 350 miles each week by car & x in summer picks fresh fruits and vegetables from x’s garden & in winter x buys fruits and vegetables trucked east from California & x writes on both sides of the paper & x uses a great deal of paper To infer ceteris paribus: x uses waste making energy

To some extent, this result is confusing, because the arguer’s purposive use of a particular language expression in reconstructing an argument now turns out be meaningless in our analysis of the argument. In this connection, we can see that Bermejo-Luque has even gone further than Freeman in this direction. As is revealed above, she assigns no role for counter-considerations within a conductive argument, and takes them to be supplemental information for assessing the warrant. Hence her reconstruction would not only blurs the distinction between conductive arguments and arguments without the mention of counter-considerations, but it would also make conductive arguments that include different counter-considerations become indistinguishable from each other. For instance, in her account, the following three arguments will be the same in reconstruction because they all have the same premises, conclusion, and most importantly, an identical warrant of “if it is true that the picture is ideal for children and it is true that it will be gone by tomorrow, then it is true that you should take your son to the movie”: [1] Although your lawn needs cutting, you ought to take your son to the movie because the picture is ideal for children and will be gone by tomorrow. [2] Although you need to finish the last part of your paper today, you ought to take your son to the movie because the picture is ideal for children and will be gone by tomorrow. [3] You ought to take your son to the movie because the picture is ideal for children and will be gone by tomorrow.

Based on these reflections on Freeman’s and Bermejo-Luque’s accounts, it is revealed that a correct reconstruction of conductive argument will need to fully respect its linguistic feature, for the distinctiveness of conductive arguments lies in the arguer’s language uses in formulating the argument, i.e., the particular way of introducing counter-considerations by “even-though” clauses. Seen in this light, it might be said that the supplementation of on-balance premise approach is more on the right track, because obviously an on-balance premise captures better the pragmatics of “even-though” clauses, i.e., counter-considerations are conceded and outweighed. However, when it comes under close examination, we could also find that such an approach remains to be unsatisfactory in characterizing the structure of conductive arguments.

246

Y. Xie

15.4 The Metaphor of Outweighing As noted in Sect. 15.2.1, the supplementation of on-balance premise approach is originated from the idea of conceiving conductive argument through the model of weighing, an idea that is first proposed by Wellman in his elaborations on conduction. Moreover, Wellman has also developed a particular understanding of the model of weighing, which has exerted great influence on the later discussions about conductive arguments. According to him, “the weighing should not be thought of as putting each reason on a scale, noting the amount of weight, and then calculating the difference between the weight of the reasons for and the reasons against”, (Wellman, 1971, p. 57) for such an interpretation will suggest a process of weighing that is “too mechanical”. Rather, as Wellman contends, “one should think of the weighing in terms of the model of determining the weight of objects by hefting them in one’s hands.”(p. 58) This way of thinking about weighing, as he has further clarified, “brings out the comparative aspect and the conclusion that one is more than the other without suggesting any automatic procedure that would dispense with individual judgment or any introduction of units of weight” (p. 58). Here it can be seen that Wellman has established a basic view on the process of weighing that highlights the comparison between the reasons for and against, but meanwhile he resists to unpack such a comparison in terms of individual weight. Therefore, weighing between the pros and the cons could only be realized by considering their collective strength, and their comparison will also become a simple one-time task, as Wellman has explained metaphorically, just like hefting two things in one’s two hands. Apparently, it is this particular Wellmanian view that predicts the supplementation of on-balance premise approach proposed later by Blair and Hansen, and as a result, it has also shaped their understanding on the structure of conductive arguments. As is indicated in Sect. 15.2.1, an on-balance premise states that the supporting considerations, when taken together, outweigh the counter-considerations taken together. The presence of that premise has claimed an outweighing relation that is understood exactly in terms of the collective strength. And by adding such a premise in the reconstruction of conductive argument, it restores a complete mechanism of weighing that is as simple as Wellman has envisaged. However, I suspect that such a mechanism is indeed too simple to fully characterize the process of weighing underlying conductive arguments. For the first, as Wellman himself has also recognized, “the degree of support is not measurable…because there is no unit of logical force in which to do the calculation” (p. 57), therefore, the notion of collective strength remains to be unclear, and a correct understanding of the outweighing relation in the on-balance premise seems to be difficult. Accordingly, unless we have an adequate grasp of how to take considerations together and to compare them, otherwise the relation of outweighing would be understandable only as a metaphor, and its explanation would always be relying upon some other similar but more intelligible situations, for example, as Wellman has offered one, like our comparing two piles of pebbles by hefting them in hands.

15 On the Logical Reconstruction of Conductive Arguments

247

For the second, even if the notion of collective strength is manageable with a sufficient model characterizing the accumulation of degrees of support in individual reasons, an outweighing relation based merely on collective strength still oversimplifies the interactions between the reasons for and against. In general, reasons can interact with each other in a variety of ways. When reasons are taken together, sometimes their interaction takes the form of increasing or decreasing the degree of support, but sometimes not. To some extent, this phenomenon was explored by Pollock in his studies on defeasible reasoning (Pollock, 1994), and has also been well recognized in the studies on the accrual of arguments in artificial intelligence (Prakken, 2005). Here we can take Pollock’s distinction between rebutting defeater and undercutting defeater as a simple illustration. According to Pollock, “information that can mandate the retraction of the conclusion of a defeasible argument constitutes a defeater for the argument” (2008, p. 453). Accordingly, there are two kinds of defeaters that can be distinguished: “rebutting defeaters attack the conclusion of a defeasible inference, while undercutting defeaters attack the defeasible inference itself.” (p. 453) In particular, undercutting defeaters can be thought of “as giving us a reason for believing that (under the present circumstances) the truth of the premises does not guarantee the truth of the conclusion” (p. 453). For instance, “your lawn needs cutting” would be a rebutting defeater to the argument of “you ought to take your son to the movie because the movie is ideal for children”, for that information could count as a reason for denying the conclusion. However, to the same argument, “your son might be different in his taste of movie” would be a undercutting defeater, because that information gives us a reason to doubt the inference from the premise “the movie is ideal for children” to the conclusion “you ought to take your son to the movie”. It is easy to see that the notion of defeater as defined by Pollock is very close to the understanding of counter-consideration in the discussions on conduction. As clarified by Govier, counter-considerations are “points that are negatively relevant to the conclusion” (2010, p. 355), or “negatively relevant factors that count against the conclusion” (2011, p. 266). In this connection, counter-considerations in a conductive argument could also be taken as some defeaters to the argument that has the supporting reason(s) and the conclusion in that conductive argument as its own premise(s) and conclusion. Likewise, the distinction between rebutting and undercutting can also be drawn for counter-considerations, which could in turn deepen our understanding of the underlying mechanism in conductive arguments. For instance, the following two conductive arguments could then be further clarified in their different uses of counter-considerations: [1] Although your lawn needs cutting, you still ought to take him to that movie because it is ideal for children and will be gone by tomorrow. [2] Although your son might be different in his taste of movie, you still ought to take him to that movie because it is ideal for children and will be gone by tomorrow.

As we have already explained above, here in the first argument by mentioning the counter-consideration the arguer is acknowledging some concern that would lead directly to the falsity (or the opposite) of the conclusion. But in the second

248

Y. Xie

argument, the arguer is conceding some concern that, by itself, does not challenge the correctness of the conclusion, and more importantly, it is a concern that is relevant only to one of the supporting reasons (i.e., that movie is ideal for children), and would only attacks the inference from that reason to the conclusion. Furthermore, in light of this parallel between defeaters and counterconsiderations, we could also find that an outweighing relation based merely on collective strength now becomes inadequate in characterizing the ways of arguing in conduction. Basically, such an outweighing relation can only capture the process of weighing in those conductive arguments that include counter-considerations of a rebutting type. More specifically, an outweighing relation makes sense only when we are considering two (sets of) reasons that support respectively conflicting conclusions (A and not-A), because it is only in this situation that we are going to take into account their own weight (or degree of support) and to determine which one is stronger, i.e., which one is rebutted by the other. Moreover, it could also be contended that a comparison of weight is meaningful only between reasons that support the same conclusion, or reasons that support conflicting conclusions, because the weight (or degree of support) of a reason is indeed a relative notion, i.e., a reason has a weight m in support of A. In view of this relativity, it is easy to see that only when two reasons are both in support of A, or respectively in support of A and notA, their weights could be compared. But if one reason is in support of A while the other in support of B, their weights, even though being both measured in degree, are not comparable, hence to speak about one of them outweighs the other would be misleading and inappropriate. For this reason, it is now clear that an outweighing relation based on collective strength is not appropriate for understanding those conductive arguments that include counter-considerations of a undercutting type, for the weight of an undercutting counter-consideration is in support of neither the conclusion nor its opposite, thus it cannot be compared with that of the supportive reasons, no matter individually or collectively. Therefore, to conclude, the metaphor of outweighing cannot capture the process of weighing in conductive arguments in a comprehensive manner, hence the supplementation of on-balance premise approach to their reconstruction remains to be immature, if not misleading.

15.5 The Role of Counter-Considerations in Conductive Arguments In spite of the simplification discussed in last section, there is still some other doubt that could be raised in regard to the adequacy of the supplementation of on-balance premise approach. As is revealed before, an on-balance premise states explicitly that the reasons for the conclusion have outweighed the counter-considerations in justifying the conclusion. Apparently, by formulating such a premise, all the counterconsiderations in a conductive argument have been regarded as playing a substantial role in justifying the conclusion. In other words, it is presumed that the arguer has

15 On the Logical Reconstruction of Conductive Arguments

249

taken counter-considerations as points relevant to the establishment of the conclusion, hence she mentioned them in the argument for logical concerns, in such a way that their justificatory force is explicitly acknowledged, but then judged as insufficient to undermine her drawing the conclusion. However, such a presumption might be doubted as simply resting on an inferential leap taken from argumentative practice to our theorizing, for even though the arguer does have offered some counterconsiderations in her argument, it remains to be unclear whether she has assigned for them a relevant role in justifying the conclusion. Therefore, if the supplementation of on-balance premise approach is to be justified, there is still a need to determine the role of counter-consideration, and thereby to establish such a presumption in a proper way. In a recent paper, Hansen (2011) has provided a detailed discussion about the role of counter-considerations in conductive arguments, and has offered, to some extent, a defense for the above presumption. For the first, he has rightly realized that such a presumption is just a hypotheses that is competing with several other alternatives (p. 40). Then, Hansen has considered six different possibilities in regard to the role of counter-considerations, focusing particularly on their weakness in characterizing the argument structure of conduction. And his assessment is summarized in the following table (p. 40):

Hypotheses

Response

1

Counter-considerations are premises

No. Only statements that are positively relevant to the conclusion can be premises of an argument

2

Counter-considerations are background knowledge

No. They are mentioned and considered in the argument

3

Counter-considerations are qualifications about the conclusion

No. It is the reasoning for the conclusion, not the conclusion, that is qualified

4

Counter-considerations are the anti-premises of a counter-argument

No. A BC-argument is one argument, not two

5

Counter-considerations and premises are both relevant to the conclusion

Yes. But there is more to a BC-argument than identifying reasons for and against the conclusion

6

Counter-considerations are said to be outweighed in an on-balance premise

OK. But this means giving up the thesis that BC-arguments have convergent premises

As Hansen has argued, the last hypotheses “overcomes the shortcomings of the other possibilities we looked at” (p. 39), thus should be taken as a better way to analyze the structure of conductive arguments. It is better because, most importantly, it captures well the particular form of reasoning that is unique to conduction, i.e., “counter-considerations are not just stated without resolution along with the reasons for the conclusion” (p. 39), but have been judged as outweighed by them. Clearly, such an unique form of reasoning implies that counter-considerations are indeed relevant points that have a bearing on the establishment of conclusion, as

250

Y. Xie

Hansen has emphasized, they have been considered and outweighed. But how do we know that this form of reasoning is unique for conduction? In answering this question, Hansen resorts to Govier’s claim that “a person who explicitly acknowledges counter-considerations and nevertheless still claims that her conclusion is supported by positively relevant premises is committed to the judgment that the positively relevant premises outweigh the counter-considerations” (Govier, 2005, p. 397). This claim was made, in turn, by Govier based on her analysis of the “words serving to introduce counter-considerations”, such as “although” and “even though” (see Govier, 2010, pp. 355–357). So, here once again we have returned back to the linguistic feature of conductive arguments. I think Govier is definitely right in revealing that, by using those concessive expressions to introduce counter-considerations in her argument, the arguer has acknowledged them, and is also committed to a judgment that they are less important (or, in her term, outweighed, i.e., lighter in weight) if compared with supporting reasons that are offered in the argument. However, such a general analysis about the pragmatics of language indicators of the even-though type, still, cannot establish the presumption that is at work in the supplementation of on-balance premise approach, because the arguer’s acknowledgement of a counter-consideration, and her commitment of its being less important as well, cannot determine the way in which the arguer is mentioning it in a particular argument. In other words, it remains to be possible that the arguer, when introducing a counter-consideration in her argument, doesn’t accept it as a point pertaining to the establishment of her conclusion in that argument. Hence, it is not so much the particular concessive way of mentioning that uncovers the specific role of counter-considerations within an argument, rather, such a role should be decided in accord with the arguer’s communicative intent in mentioning them. In light of this, now it might be easier to see that it is indeed misleading to simply presume that counter-considerations mentioned in a conductive argument are playing a logical role in justifying the conclusion. To illustrate, let’s compare the following two examples: [1] Although your lawn needs cutting, you still ought to take him to that movie because it is ideal for children and will be gone by tomorrow. And it goes without saying that you can simply ask someone to cut the grass for you, and even if there comes a warning or fine ticket, it’s really not a big deal when compared with your son’s happiness. [2] Although your lawn needs cutting, you still ought to take him to that movie because it is ideal for children and will be gone by tomorrow.

In the first example, apparently the arguer not only mentions a counterconsideration but also tries to rebut it. In particular, she has tried to remove its justificatory power in undermining her conclusion in the argument, by offering a reason against it and by indicating its being outweighed. In doing so, the arguer’s communicative intent for mentioning the counter-consideration has also become obvious, that is, to refute it and thus defend (or consolidate) the argument. So here we have truly encountered an argument in which it is clear that arguer has mentioned a counter-consideration, and has also taken it as a point pertaining to the establishment of the conclusion in her argument. And the conclusion is reached by her taking

15 On the Logical Reconstruction of Conductive Arguments

251

together both supporting reasons and a counter-consideration, and through a process of weighing. By comparison, the second argument is much simpler, and most importantly, it is different for containing less information. However, when we assign a logical role for the counter-consideration in the second argument and then reconstruct it by supplementing an on-balance premise for it, we are treating these two examples as exactly the same in their underlying mechanism of arguing. So, here the next question would be: can we treat them in that way? Or, in other words, could this treatment be justified? I think such a treatment could only be justified in two different ways. For the first, it will be reasonable to do so if their difference does not matter in our understanding of them. As is indicated, the second argument is different from the first for containing less information. More specifically, it lacks a refutation to the counterconsideration, and it also ignores to offer an explanation about how, or why, the counter-consideration is outweighed. These two things, however, not only are helpful in indicating that the arguer is drawing the conclusion by weighing reasons for and against, but they also constitute the most substantial parts of her process of weighing. Therefore, when we take these two examples as arguing in the same way, we are just putting the second arguer into an odd situation where she is truly presenting an argument that reaches the conclusion by weighing the pros and the cons, but meanwhile she simply ignored to provide the most important information regarding that process of weighing. If this is exactly the case, then to a great extent the arguer is violating the cooperative principle, a principle that demands participants in communication to make their conversational contribution as informative as is required (see Grice, 1989, 26). The second way to justify our treating these two examples in the same manner will be by means of empirical evidence, i.e., to verify that in real argumentative practices they do have been recognized as the same in their underlying mechanism of arguing. To be more specific, if it is simply an empirical findings that the arguers who present an argument in the format of the second example really are aiming to draw the conclusion in a way of weighing between supporting reasons and counterconsiderations, then we certainly should understand their arguments accordingly in that way. However, to my knowledge, there is no such kind of empirical evidence reported yet. Besides, I have once conducted a study that was designed from the perspective of the recipient of argument, in order to investigate what is the possible role of counter-considerations recognized by the addressee of the argument. And it is reported that the recipients of conductive arguments have hardly understand counterconsiderations as playing a logical role, and the mention of counter-considerations in a conductive argument is more likely to be recognized as some effort to give the argument a better appearance, or, to adapt to the audience for establishing communion and gaining adherence (Xie, 2017, pp. 6–8). Thus it can be seen neither of these two ways is successful till now. Although I am not completely pessimistic towards the justification of the same treatment for the above two examples, here I would also like to point out another alternative point of view from which we would just treat them in the opposite way, i.e., highlighting that they are indeed different. And I think from such a point of view we could also come

252

Y. Xie

up with a different account in regard to the mechanism of arguing in conductive arguments. In brief, given the fact the second argument is different from the first one in its lacking of the most important parts in the process of weighing, we could also fully respect that difference and then resist to regard it as arguing by weighing the pros and the cons. In this connection, a conductive argument is not to be conceived in terms of an underlying mechanism of weighing, for the counter-considerations mentioned in it now become points that are not truly outweighed in that argument. In other words, when the arguer introduces some counter-considerations in a conductive argument, she does acknowledge them, but she does not take them to be relevant in her justification of the conclusion in that particular argument, for there she has intentionally presented them in a non-refutational way, with no attempt at all to discuss or to remove them. Moreover, such a characterization could also be justified by replying upon the pragmatics of language indicators like “although” and “even though”, because, as we have indicated before, introducing a counter-consideration in such a particular concessive way does not necessarily mean that the arguer accepts it as a point relevant to the establishment of the conclusion. Therefore, I think this particular account of conductive argument could also count as an alternative hypotheses that needs to be examined if the supplementation of on-balance premise approach is to be finally defended.

15.6 Conclusion In this paper I provide a critical discussion on the logical approaches to the reconstruction of conductive arguments. The supplementation-of-on-balance-premise approach can be traced back to Wellman’s idea of conceiving conductive arguments in terms of a mechanism of weighing. And there are two different accounts developed respectively by Blair and Hansen. They both take the presence of counter-considerations as indicating that the conclusion is justified by weighing, hence propose to reconstruct conductive arguments by adding an on-balance premise which makes it explicit that the reasons for the conclusion have outweighed those counter-considerations. The warrant-reformulation approach is developed by Freeman and Bermejo-Luque in light of the Toulmin model. Although they both regard counter-considerations in conductive arguments as being related to the notion of warrant, they take different ways to fit them into the model. Freeman includes counter-considerations as conjunctions in the premise of the warrant, while Bermejo-Luque takes them as a separate unit that only has some value in assessing the plausibility of warrant. It is then argued that the warrant-reformulation approach has failed to give full credit to the unique linguistic feature of conductive arguments, hence the proposed ways of reconstruction simply make conduction trivialized as an argument type, and indistinguishable from many other arguments. Moreover, it is revealed that the supplementation of on-balance premise approach, though captures better the pragmatics of “even-though” clauses, relies upon a metaphor of outweighing that oversimplifies the process of weighing underlying conductive arguments. Besides, the formulation

15 On the Logical Reconstruction of Conductive Arguments

253

of an on-balance premise also presumes a logical role for counter-considerations, but such a presumption is misleading, because it simply disregards the difference between two distinct argumentative phenomenon in a way that is still in need of sufficient defense. Therefore, to conclude, the current logical approaches to the reconstruction of conductive arguments remain to be unsatisfactory, there is still a lot of work to do if we want to develop them into better accounts. Acknowledgements The work in this paper is supported by the National Social Science Fund of China (18ZDA033).

References Adler, J. E. (2013). Are conductive arguments possible? Argumentation, 27(3), 245–257. Bermejo-Luque, L. (2011). Giving reasons: A linguistic-pragmatic approach to Argumentation Theory. Dordrecht: Springer. Bermejo-Luque, L. (2017). The appraisal of conductions. In S. Oswald & D. Maillat (Eds.), Argumentation and inference (pp. 1–18). London: College Publications. Blair, J. A. (2016). A defense of conduction: A reply to Adler. Argumentation, 30(2), 109–128. Blair, J. A. (2017). In defence of conduction: Two neglected features of argumentation. In S. Oswald & D. Maillat (Eds.), Argumentation and inference (pp. 29–44). London: College Publications. Blair, J. A., & Johnson, R. H. (Eds.). (2011). Conductive arguments: An overlooked type of defeasible reasoning. London: College Publications. Freeman, J. B. (2011). Evaluating conductive arguments in light of the Toulmin model. In J. A. Blair & R. H. Johnson (Eds.), Conductive arguments: An overlooked type of defeasible reasoning (pp. 127–144). London: College Publications. Govier, T. (1979). Carl Wellman’s challenge and response. The Informal Logic Newsletter, 2(2), 10–15. Govier, T. (1987). Problems in argument analysis and evaluation. Dordrecht: Foris Publications. Govier, T. (2005). A practical study of argument (6th ed.). Belmont CA: Thomson/Wadsworth. Govier, T. (2010). A practical study of argument (7th ed.). Belmont CA: Wadsworth, Cengage Learning. Govier, T. (2011). Conductive arguments: Overview of the symposium. In J. A. Blair & R. H. Johnson (Eds.), Conductive arguments: An overlooked type of defeasible reasoning (pp. 262–276). London: College Publications. Grice, H. P. (1989). Logic and conversation. In H. P. Grice (Ed.), Studies in the way of words (pp. 22–40). Cambridge, MA: Harvard University Press. Hansen, H. V. (2011). Notes on balance-of-consideration arguments. In J. A. Blair & R. H. Johnson (Eds.), Conductive arguments: An overlooked type of defeasible reasoning (pp. 31–51). London: College Publications. Juthe, A. (2018). Reconstructing Complex Pro/Con Argumentation, Argumentation, online first. Pollock, J. (1994). Justification and defeat. Artificial Intelligence, 67, 377–407. Pollock, J. (2008). Defeasible Reasoning, In J. Adler & L. Rips (Eds.), Reasoning: Studies of Human Inference and its Foundations (pp. 451–470), Cambridge University Press. Prakken, H. (2005). A study of accrual of arguments, with applications to evidential reasoning. Paper presented at International conference on artificial intelligence and law (ICAIL’05), Bologna, Italy. van Laar, J. A. (2014). Arguments that take Counter-considerations into Account. Informal Logic, 34(3), 240–275.

254

Y. Xie

Wellman, C. (1971). Challenge and response: Justification in Ethics. Carbondale & Edwardsville: Southern Illinois University Press. Xie, Y. (2017). Conductive argument as a mode of strategic maneuvering. Informal Logic, 37(1), 2–22.

Chapter 16

The Legitimacy of Conductive Arguments: What Are the Logical Roles of Negative Considerations? Yanlin Liao

16.1 Introduction In 1971, the American philosopher Carl Wellman claimed that there is a distinct type of argument—“conduction” (also known as conductive arguments), in his book Challenge and Response: Justification in Ethics. He contends that conduction is an argument which is neither deductive or inductive (Wellman, 1971, pp. 4–51): Deduction is that form of reasoning in which the claim is made that the conclusion follows necessarily from the premises… (p. 4) By “induction” I mean that sort of reasoning by which a hypothesis is confirmed or disconfirmed by establishing the truth or falsity of its implications. To show that the consequences of some hypothesis are true is to provide evidence for its acceptance; to show that one or more of its consequences are false is to refute it. It is this sort of reasoning, so important to science, to which I refer by the word “induction.”… (p. 32) When one has considered all the deductive and inductive arguments in ethics, there remain others to be taken into account. A few ethical arguments that fall into neither of the two traditional categories are the following: You ought to take your son to the circus because you promised. This is a good book because it is interesting and thought-provoking. Although he is tactless and nonconformist, he is still a morally good man because of his underlying kindness and real integrity. In most cases this sort of reasoning derives its conclusion from a variety of premises each of which has some independent relevance. Typically, although by no means always, several reasons are given in such arguments; and in those cases where a single reason is advanced there are others which might have been given as well. Since what is characteristic of this sort of reasoning is the leading together of various considerations, it seems appropriate to label it “conduction.” (p. 51)

Y. Liao (B) Modern Logic and Application Institute/Department of Philosophy, Nanjing University, Nanjing, Jiangsu, China e-mail: [email protected] © Springer Nature Switzerland AG 2020 F. H. van Eemeren and B. Garssen (eds.), From Argument Schemes to Argumentative Relations in the Wild, Argumentation Library 35, https://doi.org/10.1007/978-3-030-28367-4_16

255

256

Y. Liao

It would be beyond the scope of this paper to discuss the tenability of the taxonomy of “deduction-induction-conduction” presented by Wellman. Here we just need to acknowledge that Wellman thinks that conduction, which is claimed to be neither deductive or inductive, is a distinct type of argument (reasoning).1 Moreover, he elaborates the “nature” of conduction (p. 52): Conduction can best be defined as the sort of reasoning in which (1) a conclusion about some individual case; (2) is drawn non-conclusively; (3) from one or more premises about the same case; (4) without any appeal to other cases.

Due to the different composition of premises, there are three “patterns” of conduction (pp. 55–57): The first pattern is exhibited by any conductive argument in which a single reason is given for the conclusion. Examples might be “you ought to help him for he has been very kind to you” or “that was a good play because the characters were so well drawn.”… In the second pattern of conduction several reasons are given for the conclusion. Examples are “you ought to take your son to the movie because you promised to do so, it is a good movie, and you have nothing better to do this afternoon” or “this is not a good book because it fails to hold one’s interest, is full of vague description, and has a very implausible plot.”… The third pattern of conduction is that form of argument in which some conclusion is drawn from both positive and negative considerations. In this pattern reasons against the conclusion are included as well as reasons for it. For example “in spite of a certain dissonance, that piece of music is beautiful because of its dynamic quality and its final resolution” or “although your lawn needs cutting, you ought to take your son to the movies because the picture is ideal for children and will be gone by tomorrow.”…

However, recent discussions on conduction mainly focuses on the third pattern of conduction (also called “pro-con arguments” or “balance-of-considerations arguments”). For example, J. A. Blair and R. H. Johnson published a compilation of papers, called Conductive Argument: An overlooked Type of Defeasible Reasoning in 2011. Most of the papers in the compilation discussed relevant issues on the third type of conduction.2 In 2013, Adler argued that conduction is “impossible” because the definition of conduction implies two “incompatible” claims, which posed a challenge to the legitimacy of conduction. From then on, the legitimacy of conduction had been the most important and controversial issue in the field of conduction. Blair (2016), Possin (2016), Xie & Xiong (2013) and Xie (2017) actively participated in the debate on this issue in recent years. Unfortunately, researchers have been far from reaching an agreement on the legitimacy of conduction. As Possin said, “After decades of debate, the nature of conductive arguments has become progressively less clear.” (2016, p. 564) In this paper, I first try to clarify and review the disputes on the legitimacy of conduction and then present a new approach to justify the legitimacy of conduction.

1 Wellman

seems not to make a distinction between “reasoning” and “argument” in his book. term “conduction” or “conductive arguments” hereafter this paper merely refers to the third type of conduction (i.e., pro-con arguments).

2 The

16 The Legitimacy of Conductive Arguments: What …

257

16.2 “Adler’s Problem” and Its Variant If you go through the recent debate on the legitimacy of conduction, you will probably get lost in the complicated and vague disputes. I will explicitly reconstruct the debate to show what the legitimacy crisis of conduction is. First of all, we can present a brief definition of conductive arguments based on Wellman’s above description: An argument is a conductive argument if and only if: (1) it satisfies four characteristics of conduction mentioned by Wellman; (2) its premises include positive and negative (counter) considerations.

However, Adler (2013, p. 249) argues that the definition of conduction implies two “incompatible” claims which make conduction “impossible.” The claim that the definition of conduction makes conduction impossible will be called “Adler’s problem” in this paper. The two incompatible claims implied by the definition of conduction are below (Ibid.): (I) The counter-considerations are irreducible or ineliminable—the conclusion, C, is reached without nullifying the counter-considerations. (II) C is accepted as true, which issues in belief. C is detached from these premises (reasons) so that what is accepted or believed is C itself.

These two claims presented by Adler are not understandable enough because of some confusing terms he used in the claims (e.g. “irreducible,” “ineliminable,” “nullifying,” “detached”). In order to evaluate Adler’s argument, Blair (2016, pp. 111–114) explicitly interprets the two claims above based on a careful examination on Adler’s text, which is very helpful for us to understand what “Adler’s problem” is. For the interpretation of the first claim, Blair writes (Ibid, p. 112): Adler says: “The argument yields belief in the conclusion. However, this latter claim conflicts with the characterization of conductive arguments as inconclusive, since undermining reasons remain viable” (2013, p. 248). From these passages, we can infer that Adler takes the “inconclusive” or “non-conclusive” property attributed to conductive arguments to in their counter-considerations continuing to carry force and to weaken the argument even after the conclusion has been drawn.

This is to say, Adler thinks that conductive arguments are “non-conclusive” in the sense that counter-considerations still weaken the argument even after the conclusion has been drawn. In this sense, the counter considerations are “irreducible” (i.e., they still weaken the argument). Moreover, for the interpretation of the second claim, Blair says (Ibid., p. 112): He [Adler] also allows that there can be conductive-type arguments in which the conclusion is qualified by some such term as ‘prima-facie’ or ‘probably,’ but he notes that in such cases the conclusions are not detachable, since no “all things considered” conclusion is drawn. The cases he is interested in are those in which a conclusion is drawn without qualification, on the basis of a consideration of all the available epistemically relevant representative pro and con considerations (Adler, 2013, p. 253).

258

Y. Liao

For Adler, a conclusion “detached” from the premises means that a conclusion is drawn without any qualification. So, according to Blair’s clarification, Adler’s problem can be rephrased as below: (Ia )

Conductive arguments are “non-conclusive,” so the negative considerations remain viable (continue to weaken the argument) even after the conclusion has been drawn. (IIa ) The conclusion of conductive arguments should be without any qualification (e.g., “prima-facie,” “probably”) Adler seems to thinks that (Ia ) and (IIa ) are explicitly conflicting (“incompatible”). So, Adler’s problem can make conduction impossible as long as these two claims both can be accepted by the proponents of conduction. In this connection, Adler (2013, p. 249) contends that (Ia ) is obviously acceptable for the proponents, because it is the “the standard view” of conductive arguments. However, he thinks that (IIa ) is “the crux of the argument” that requires defense. Therefore, he makes much effort to justify (IIa ) from different aspects (Ibid., p. 250): First, as the illustrations—particularly those of Wellman and Hitchcock—show, ‘probably’ and ‘prima facie’ conclusions are uncharacteristic of Conductive Arguments. Further illustrations are offered below. Each reads most naturally as drawing an unqualified conclusion. Second, these illustrations fit with argument as primarily aiming to discern or advance and settle new or interesting or important truths, that are worth believing for ourselves or for our audience. They increase our information and expand our corpus of beliefs. Third, only arguments that end with detached, unqualified conclusions represent an end of inquiry. Only then are the arguers and inquirers in a position that is appropriate to guide further judgments and action.

Interestingly, Blair completely agrees what Adler defends for (IIa ), but does not think that (Ia ) is acceptable. Blair (2016) points out that the problem with (Ia ) is that Adler misunderstands the meaning of “non-conclusive” or “inconclusive.” For Adler, “non-conclusive” means that “undermining reasons remain viable” (2013, p. 248). After reviewing Wellman, Govier and Pinto’s text regarding conductive arguments, however, Blair clarifies the term “non-conclusive” (2016, p. 116): It seems clear that ‘non-conclusive’ is used in different ways by these proponents of conductive Adler does misunderstand the arguments, on the one hand, and by Adler in his criticism of the possibility of conductive arguments, on the other. The former mean that conductive arguments are non-conclusive in the sense that their premises do not deductively entail their conclusions—that they are defeasible; but Adler takes them to mean that such arguments are non-conclusive in the sense that their conclusions are not detached from their premises, but always remain qualified by the acknowledged counter-considerations. Adler simply mistakes what those theorists mean by ‘non-conclusive.’

So, (Ia ) cannot be accepted by the proponents of conduction because Adler misunderstands the term “non-conclusive” the proponents used. More importantly, Blair (Ibid.) notices that Adler holds that the conclusions of defeasible arguments can be “detached.” Given this, if Adler could correct his misinterpretation of the term “nonconclusive,” Blair seems to think that the alleged Adler’s problem would be revised as follows:

16 The Legitimacy of Conductive Arguments: What …

259

(Ib ) Conductive arguments are “non-conclusive,” so they are defeasible arguments. (IIb ) The conclusion of conductive arguments should be without any qualification (e.g., “prima-facie,” “probably”). Since Adler holds that the conclusions of defeasible arguments can be without any qualification, (Ib ) and (IIb ) are not conflicting at all. Thus, Adler’s problem is resolved in this way. That is why Blair says that “there seems to be no incompatibility between Adler’s position and that of the proponents of conductive arguments.” (Ibid.) Unfortunately, Adler’s problem has not yet been resolved so far. Although the proponents of conduction might not accept Adler’s understanding of the term “nonconclusive,” Xie and Xiong believe that they would accept the point that “the negative considerations remain viable even though they are outweighed by positive considerations” (2013, pp. 3–4, emphasis in original): Regardless of the fact that Adler does misunderstand the others’ uses and meanings of inconclusive, let’s try to think the issue in another way around: whether the others have indeed attributed conductive arguments with a characterization that Alder has referred to by his wrong use of the term inconclusive? We believe they have, since for almost all of them, counter-considerations, though outweighed, are still seen to be remaining viable, and are treated as premises providing reasons relevant to the strength of argument. It is nearly the same as the property that Adler has attributed to conductive arguments with “inconclusive.”

I agree with Xie and Xiong. For the proponents of conduction, the property that negative considerations remain viable makes conduction be viewed as a distinct type of argument—conductive arguments. In other words, the property plays a critical role in the legitimacy of conductive arguments. So, I believe that they will support the property. Given this, we can give up using the term “non-conclusive” and modify (Ia ) to (Ic ) which will be accepted by the proponents: (Ic ) In conductive arguments, negative considerations remain viable (continue to weaken the argument) even after the conclusion has been drawn. By doing this, Adler’s problem becomes the variant of Adler’s problem: c

(I )

In conductive arguments, negative considerations remain viable (continue to weaken the argument) even after the conclusion has been drawn. (IIc ) The conclusion of conductive arguments should be without any qualification (e.g., “prima-facie,” “probably”). It is clear that (I) and (Ic ) both express that “negative considerations remain viable.” So, for Adler, the variant of Adler’s problem will still make conduction impossible. Moreover, the variant of Adler’s problem can be further summarized as a question: is it possible that the conclusions of conductive arguments are unqualified while the negative considerations remain viable? If the answer to the question is “yes,” then conduction will be possible, which means that the legitimacy of conduction will be justified. Otherwise, the conduction will be impossible, which means that the legitimacy will be unjustified. There are two different answers to the question, which I will call “the rhetorical solution” and “the logical solution” respectively.

260

Y. Liao

16.3 The Rhetorical Solution Versus the Logical Solution The rhetorical solution was first raised by Xie & Xiong (2013) and was systematically presented by Xie (2017). Xie explains that what the rhetorical solution is: However, it is worth noting that neither the conveying of these conventional implications, nor the achieving of those practical effects, would necessarily require the arguer to take counterconsiderations as reasons or premises pertaining to the establishment of her conclusion. Alternatively, it is possible to interpret the arguer’s communicative intention as to strengthen her argument in a way that does not enhance its justificatory power but increases its persuasive effect, i.e., makes the argument more likely to induce the adherence of the audience. I would like to call this interpretation a rhetorical perspective on conduction, simply because it regards the mention of counter-considerations as some effort aiming for rhetorical concerns to achieve better persuasiveness (2017, pp. 5–6, emphasis in original). Roughly speaking, weighing and balancing is a process of thinking in which we critically inquire into an issue, trying to consider both the pros and cons in order to reach a certain (reasonable) view. The use of a conductive argument, on the other hand, is an act of arguing by which we intend to persuade some other to accept a particular view that we have already reached and currently hold…In this connection, a conductive argument, as is presented when put to use, is merely a short-circuited retrospective reconstruction of a finished process of weighing and balancing, without the inclusion of the outweighing-relation and its justification. Therefore, it is misleading to characterize conduction simply as a special structure in which both affirmative and negative points bearing on the conclusion are explicitly collected, weighed and balanced (Ibid., p. 18, emphasis in original).

Xie makes a distinction between thinking and arguing. The negative considerations are only viable during thinking, which is a stage before arguing. As soon as entering the stage of arguing, the negative considerations are not viable anymore because the particular view has been already reached and currently hold. In other words, (Ic ) is simply not true because negative considerations are not viable in conductive arguments (i.e., arguing) though they are viable during the process of thinking. Moreover, the reason why conductive arguments consist of negative considerations is that the negative considerations can increase “persuasive effects” of the conductive arguments. He argues the pure rhetorical roles of negative considerations of conduction in different aspects (Ibid., pp. 5–13): Firstly, the negative considerations often follow the adverbials of concession, like “even though,” “although” and “notwithstanding.” From the theory of Gricean conventional implication, these adverbials could serve to convey that “these reasons against the conclusion are outweighed” or that “the speaker has taken account of not just favorable considerations, but unfavorable ones as well.” In a word, the mention of negative considerations is to make the argument appear to be more persuasive, thereby making the conclusion appear to be more solid. Second, Xie conducts an empirical study about the roles of negative considerations in conduction. 401 Subjects were asked to compare two arguments: argument A consists of one conclusion and two supporting reasons (considerations) while argument B is made from the argument A by adding two more negative considerations for the

16 The Legitimacy of Conductive Arguments: What …

261

conclusion. Most of the subjects (80.7%) think that argument B is more persuasive.3 Importantly, 89.3% of subjects do not recognize negative considerations as reasons or premises. They give some explanation of the role of negative considerations, such as “made the argument appear to be well-considered,” “serve as a foil stressing the significance of the positive reasons,” “retreating in order to advance.” The result provides empirical support for the rhetorical perspective on conduction. Third, Xie gives a detailed account of the rhetorical mechanism of negative considerations of conduction from “topical potential,” “audience demand” and “presentational devices,” which are the conceptual tools from a Pragma-dialectical framework—“strategic maneuvering.” To sum up, the variant of Adler’s problem will be able to be resolved by the rhetorical solution in this way: the negative considerations are the purely rhetorical tools, which means that they do not have the logical functions (i.e., continue to weaken the argument), so that (Ic ) is not true. In other words, the rhetorical solution resolves the variant of Adler’s problem by denying the problem per se (i.e., denying that negative considerations remain viable). Then, does the rhetorical solution help to justify the legitimacy of conduction? It does not. On the contrary, it destroys the legitimacy of conduction because it denies that the negative considerations have the logical function. As Blair commented, if weighing and balancing is only taking place before arguing, then conduction will be regarded as a distinct pattern of reasoning rather than argument (2016, p. 123). As I suggested before, in addition to the rhetorical solution, there is a logical solution to the variant of Adler’s problem. As a response to the rhetorical solution, Blair (2016) gives a brief account of the logical solution. Blair claims that negative considerations do not only appear in one’s reasoning but also can “legitimately” appear in arguments (2016, pp. 123, emphasis in original). However, overridden counter-considerations that weighed against an argument’s ultimate conclusion in one’s reasoning (the deliberation that culminated in embracing the conclusion) can legitimately appear in arguments (the reasons offered to convince others of the conclusion) so long as their role is to serve as evidence of the fact that there are conflicting considerations in play. Indeed, it is a legitimate and distinctive feature of conductive arguments that the arguer treats counter-considerations not as objections that must be refuted if the main conclusion is to be defended, but rather as objections that are granted or accepted, and which therefore must be outweighed by the pro-considerations if the main conclusion is to be successfully defended.

It is clear that Blair does not want to resolve the variant of Adler’s problem at the cost of sacrificing the legitimacy of conductive arguments. Instead, he points out that the negative considerations do not function as purely rhetorical tools, but has their logical roles (functions) in conduction (Ibid., p. 124): The mention of the con considerations indicates that the argument is weaker than it would be thought to be if only the pro considerations had been mentioned. For example, “she deserved first prize in the literary contest because, although her prose style did not stand out, 3 Subjects

are university students from various majors, but all of them are taking an introductory course of logic when they take this survey.

262

Y. Liao

her plot was so ingenious and her characters so interesting” is not as strong a case for first as, “She deserved first prize in the literary contest because her plot was so ingenious and her characters so interesting.” The conclusion, that she deserved first prize, is detached; but she did not deserve first prize simply because of her ingenious plot and interesting characters. She deserved it because of those features of her story plus the fact that their presence was needed to outweigh the ordinariness of her prose style, and succeeded in doing so.

Although Blair does not explicitly explain why the first argument of the example is weaker than the latter argument of the example, I think his evaluation is intuitively correct. I will elaborate on this point with the same example: Ar1. Her plot was so ingenious and her characters were so interesting, she deserved first prize in the literary contest. Ar2. Although her prose style did not stand out, her plot was so ingenious and her characters were so interesting; therefore, she deserved first prize in the literary contest. It is very reasonable to think that the prose style decides whether the article is special enough or not, which is one of the critical criteria of the article evaluation. Probably because of this, Blair claims that Ar2 is weaker than Ar1. By doing this, the variant of Adler’s problem can be resolved in a logical way: if a certain nonconductive defeasible argument which only has the positive considerations and the conclusion (e.g., Ar1) introduces negative considerations, the new argument (e.g., Ar2) will become a weaker argument. It is possible that the conclusions of conductive arguments are unqualified while the negative considerations remain viable because the logical solution cleverly transfers the qualification of the conclusions to the qualification of the argument strength (i.e., the supporting degree between the premises and the conclusions). Xie comes up with a criticism on the logical approach. He argues that the weakness of conductive arguments does not lie in the existence of negative considerations, but some kind of incompleteness (2017, p. 17): “arguing conductively has arbitrarily presumed as acceptable a working outweighing-relation that has played too fundamental a role in justifying the conclusion, hence cannot be left assumed and unexamined.”4 However, I do not agree with his criticism. Because I think that the negative considerations and the outweighing-relation of conductive arguments are indivisible. Only if the negative considerations exist, the working outweighing-relation is needed to be assumed. In other words, the root of the relative weakness of conductive arguments still lies in the logical roles of negative considerations. It is unconvincing to say that the relative weakness of conductive arguments has nothing to do with the existence of negative considerations. I have argued that the logical solution can resolve the variant of Adler’s problem. So far, can we conclude that the legitimacy of conduction is fully justified? Unfortunately, we cannot conclude. 4 The outweighing-relation refers to an on-balance premise which shows that the negative considera-

tions are outweighed by the positive considerations. For more discussion on the on-balance premise, see Hansen (2011) and Jin (2011).

16 The Legitimacy of Conductive Arguments: What …

263

16.4 The Perspective of Argument Evaluation As I stated in the previous section, it is “intuitively correct” that Ar2 is weaker than Ar1 because the negative consideration (i.e., “her prose style did not stand out”) apparently weakens the conclusion (i.e. “she deserved first prize in the literary contest”). That is exactly what Adler thinks that negative considerations are “viable,” which is the typical situation of conductive arguments. However, the judgment based on intuition is not always reliable. For example, there are the subtle situations which should not be ignored: when positive considerations are much more powerful than negative considerations (i.e., positive considerations make negative considerations completely insignificant), it is very hard to say that negative considerations weaken the argument. For instance: Ar3. Although her article was not very long, her plot was so ingenious and her characters were so interesting; therefore, she deserved first prize in the literary contest. Then, is Ar3 a conductive argument? According to the logical solution presented by Blair, the key to this question is whether the negative consideration weakens the argument. It is very reasonable to think that compared to the plot and characters of the article, the length of the article hardly influences the evaluation of the article in the literary contest. In this sense, it is hard to say that “her article was not very long” really weakens the conclusion. Therefore, it will be dubious to say that the argument strength of Ar3 is weaker than that of Ar1. In the subtle situations (i.e., positive considerations are much more powerful than negative considerations), the intuition on the argument strength will be questionable, which implies that the logical roles of negative considerations are questionable too.5 Some might doubt whether we need to consider the subtle situations because they think that the negative considerations are supposed to clearly weaken the conclusion. I think that, however, it is necessary to discuss the situation, because it is even the most typical case of conduction in some logicians’ minds. For example, James Freeman (2011, p. 28) thinks that the if we view negative considerations as premises just like positive considerations, then negative considerations will support conclusions “in a funny way.” Freeman contends that negative considerations should be viewed as “a special type of rebuttal or defeater” rather than “premises,” which are so weak that no counterrebuttal would be needed.6 In Freeman’s view, the alleged subtle situations of conduction are exactly the most typical conduction. In view of the unreliability of the intuition on the argument strength, we hope to evaluate the argument strength in a systematic and precise way. So the question 5 Some might think that “her article was not very long” does weaken the conclusion slightly. However,

it is also reasonable to think that it does not matter how long an article is in a literary contest. That is why I think that the intuitive judgment on the logical force of negative considerations in the subtle situation will be so dubious and controversial. 6 “Rebuttal” or “defeater” here means that the evidence negatively relevant to the conclusion being argued. For more details about rebuttals or defeaters, see Freeman (2011, pp. 20–29).

264

Y. Liao

is: can we judge argument strength in a precise manner so that we can know if Ar3 is weaker than Ar1? Unfortunately, we cannot so far. The argument strength of defeasible arguments is like a “spectrum” (from the strongest to the weakest), which is essentially different from the concept of “validity” (all or nothing). In order to evaluate the strength of conductive arguments, theorists have come up with various methods. For instance, Govier (2013, pp. 361–363) offers a concise method of “the range of exception,” and Fischer (2011) gives a more elaborative examination of some crucial concepts mentioned by Govier’s method (e.g. “case,” “ceteris paribus”). Besides, Pinto (2011) criticizes Govier’s method and tries to evaluate the strength of conduction through the concept of “relative strength.” Nevertheless, these methods do not seem to cope with the comparison between the strength of Ar1 and Ar3. Due to the difficulty of the comparison between the argument strength of Ar1 and Ar3, the logical solution or the strategy of appealing to the strength of argument fails to justify the logical roles of negative considerations in the subtle situations of conductive arguments. By contrast, the rhetorical solution can explain well what the rhetorical roles of negative considerations are in the situations—making the argument appears to be well-considered and more persuasive. Does it mean that we have to admit that the logical roles of negative considerations cannot be well justified in some cases? In other words, do we have to admit that the legitimacy of conduction cannot be well justified in some cases? If so, it will be groundless to say that Ar3 is a conductive argument. In what follows, I will provide a new proposal for the justification of the logical roles of negative considerations. Although the proposal is created to defend the logical roles of negative considerations, it is different from the logical solution presented by Blair. As I mentioned in last section, the strength of argument refers to the supporting degree between premises and conclusions. Then, how exactly do we evaluate the argument strength? For non-conductive defeasible arguments which only have the positive considerations and the conclusions (e.g., Ar1), their argument strength only depends on that to what extent the positive considerations provide support for the conclusions. So, the criterion of argument strength of non-conductive defeasible arguments is: Cr1. To what extent the positive considerations provide support for the conclusions. It means that if and only if the positive considerations of a certain non-conductive defeasible argument provide sufficient support for the conclusion, the argument is strong. For conductive arguments (e.g., Ar2, Ar3), however, their argument strength depends on two different and independent criteria: Cr1. To what extent the positive considerations provide support for the conclusions. Cr2. To what extent the positive considerations outweigh the negative considerations. It means that if and only if the positive considerations of a certain conductive argument provide sufficient support for the conclusion, and the positive considerations far outweigh the negative considerations, then the argument is strong. From

16 The Legitimacy of Conductive Arguments: What …

265

non-conductive defeasible arguments to conductive arguments, the criteria of argument strength are changed from one item (i.e., Cr1) to two items (i.e., Cr1 and Cr2). The change in the criteria of argument strength is caused by the existence of negative considerations. For a defeasible argument with no negative consideration, Cr2 will be no need to be added into the criteria of argument strength. Surprisingly, many researchers seem to notice Cr2 only but overlook Cr1. For instance, Wellman (1971, p. 57) thinks that we should decide whether the conductive argument is “valid” (i.e., strong) by weighing the pros against the cons. Govier (2013, p. 359) says that the strength of the argument is determined by “our judgment as to whether the premises cumulate to provide enough for the conclusion to outweigh negatively relevant factors.” However, if a certain argument passes the Cr2 (i.e., the positive considerations of the argument can far outweigh the negative considerations) but fails to pass Cr1 (i.e., the positive considerations of the argument cannot provide sufficient support of the conclusion), the argument will not be a strong argument. For example: Ar4. Although Peter unexpectedly failed the math exam last time, he passed math exams any other times in school, so he had perfect academic performance in math. In Ar4, the positive consideration (i.e. “he passed math exams any other times in school”) outweighs the negative consideration (i.e. “Peter unexpectedly failed the math exam last time), but the positive consideration cannot provide sufficient support to the conclusion (i.e. “he had perfect academic performance in math”). It is obvious that there is a huge gap between getting passes in math exams and doing perfectly in math exams. Thus, Ar4 is not a strong argument. To sum up, the criteria of argument strength of conductive arguments are Cr1 and Cr2. The logical roles of negative considerations are well shown by the change of criteria of argument strength from non-conductive defeasible arguments to conductive arguments. Therefore, the legitimacy of conduction can be justified. There might be two potential objections to my new proposal for the justification of the logical roles of negative considerations. The first objection goes like this: the new proposal is expected to justify the logical roles of negative considerations of the subtle situations of conduction. However, in the subtle situations of conduction, the negative considerations must be very weak. It means that Cr2 must be well satisfied, so it is unnecessary to mention Cr2 in the evaluation of the subtle situations of conduction. In reply to this objection, I would like to make two points. First, the new proposal is expected to justify the logical roles of negative considerations of all conductive arguments rather than just the subtle situations of conductive arguments. Second, I think that it inverts the procedures and products of the evaluation of argument strength. The procedures of the evaluation of argument strength mean that what criteria we will use to evaluate the argument strength, while the products of the evaluation of argument strength mean that what results we will get after the evaluation. It makes no sense to say that it is unnecessary to mention Cr2 in the evaluation of the subtle situations of conduction because we have to first use Cr1 and Cr2 to evaluate the

266

Y. Liao

argument strength and then we can judge whether the argument can be viewed as the alleged subtle situations of conduction. The second objection goes like this: for all kinds of defeasible arguments, the criteria of argument evaluation can always be Cr1 and Cr2. We just need to notice that Cr2 will be satisfied trivially (i.e., the positive considerations vacuously outweigh the negative considerations) when we evaluate the non-conductive defeasible arguments because there is no negative consideration in the arguments at all. This is to say, it is not true that the criteria of argument strength will be changed from non-conductive defeasible arguments to conductive arguments. I agree that it might be reasonable to say that the criteria of argument evaluation are always Cr1 and Cr2, even though it seems a little strange to me. However, this objection fails to cover up the logical roles of negative considerations. For nonconductive defeasible arguments, Cr2 which is always satisfied trivially is just a vacuous criterion of the argument strength. However, for conductive arguments, due to the existence of negative considerations, Cr2 will not be a vacuous criterion anymore and becomes a substantial criterion (i.e., it is untrue that Cr2 is always satisfied trivially). Therefore, the logical roles of negative considerations can still be justified from the change from a vacuous criterion to a substantial criterion. In summary, even though the new proposal which focuses on the criteria of argument strength seems to be simple, it successfully justifies the logical roles of negative considerations by arguing that the criteria of argument strength will be changed from non-conductive defeasible arguments to conductive arguments.

16.5 Conclusion In this paper, I argue that the legitimacy of conductive arguments can be fully justified. First, the alleged impossibility of conduction raised by Adler can be resolved. The core of Adler’s challenge can be summarized as a question: is it possible that the conclusions of conductive arguments are unqualified while the negative considerations remain viable? In reply to this question, the logical solution presented by Blair contends that it is possible because if a certain non-conductive defeasible argument which only has the positive considerations and the conclusion (e.g., Ar1) introduces negative considerations, the new argument (e.g., Ar2) will become a weaker argument. The logical solution cleverly resolves the Adler’s challenge by transferring the qualification of the conclusions to the qualification of argument strength. Second, the logical roles of negative considerations can be fully justified. The legitimacy of conductive arguments cannot be fully justified by Blair’s logical solution, because Blair’s solution will be dubious in some subtle situations of conduction (e.g., Ar3). In this connection, I present a new proposal which focuses on the criteria of argument strength. From non-conductive defeasible arguments to conductive arguments, the criteria of argument strength are changed from one item (i.e., Cr1) to two items (i.e., Cr1 and Cr2). This change is caused by the existence of negative considerations, which shows the logical roles of negative considerations.

16 The Legitimacy of Conductive Arguments: What …

267

It is also important to note that though I think that negative considerations have the logical roles in conduction, I do not reject that they have the rhetorical roles at the same time. I agree with the rhetorical analysis of negative considerations given by Xie and Xiong. Also, I believe that the most interesting part of conduction is that the negative considerations have both logical and rhetorical roles. Acknowledgements The rough version of this paper was presented in the conference of ISSA 2018 in Amsterdam. I truly appreciate the helpful comments and feedback from David Godden, Harald Wohlrapp, Mark Battersby, and J. Anthony Blair. And I am also very grateful for the constructive suggestions for the earlier draft of this paper by Yun Xie of Sun Yat-sen University and Jianjun Zhang of Nanjing University.

References Adler, J. E. (2013). Are conductive arguments possible? Argumentation, 27, 245–257. Blair, J. A. (2016). A defense of conduction: A reply to Adler. Argumentation, 30, 109–128. Blair, J. A., & Johnson, R. H. (Eds.). (2011). Conductive argument: An overlooked type of defeasible reasoning. London: College Publications. Fischer, T. (2011). Weighing considerations in conductive pro and con arguments. In J. A. Blair & R. H. Johnson (Eds.), Conductive argument: An overlooked types of defeasible reasoning (pp. 86–103). London: College Publications. Freeman, J. B. (2011). Argument structure: Representation and theory. Springer Science + Business Media B.V. Govier, T. (2013). A practical study of argument (enhanced (7th ed.). Boston: Wadsworth Cengage Learning. Hansen, H. V. (2011). Notes on balance-of-considerations arguments. In J. A. Blair & R. H. Johnson (Eds.), Conductive argument: An overlooked types of defeasible reasoning (pp. 31–51). London: College Publications. Jin, R. (2011). The structure of pro and con arguments. In J. A. Blair & R. H. Johnson (Eds.), Conductive argument: An overlooked types of defeasible reasoning (pp. 10–30). London: College Publications. Pinto, R. C. (2011). Weighing evidence in the context of conductive reasoning. In J. A. Blair & R. H. Johnson (Eds.), Conductive argument: An overlooked types of defeasible reasoning (pp. 104–126). London: College Publications. Possin, K. (2016). Conductive arguments: why is this still a thing? Informal Logic, 36, 563–593. Wellman, C. (1971). Challenge and response: Justification in ethics. Carbondale, IL: Southern Illinois University Press. Xie, Y. (2017). Conductive argument as a mode of strategic maneuvering. Informal Logic, 37, 2–22. Xie, Y., & Xiong, M. (2013). Commentary on: J. Anthony Blair’s “are conductive arguments really not possible?” In D. Mohammed & M. Lewi`nski (Eds.), Virtues of argumentation: Proceedings of the 10th international conference of the Ontario society for the study of argumentation (pp. 1–6). Windsor, ON: OSSA.

Chapter 17

Deploying Machine Learning Classifiers for Argumentative Relations “in the Wild” Oana Cocarascu and Francesca Toni

17.1 Introduction Argument Mining (AM) aims at automatically identifying arguments and components of arguments in text, as well as at determining the relations between these arguments, on various annotated corpora using machine learning techniques (Lippi & Torroni, 2016). We focus on Relation-based Argument Mining (RbAM) (Carstens & Toni, 2015) to determine argumentative relations of attack and support between any two texts, a step towards extracting, from text, Bipolar Argumentation Frameworks (BAFs) (Cayrol & Lagasquie-Schiex, 2005), namely graphs whose nodes are arguments and with two types of edges (representing attack and support relations). These graphs can then be deployed to support a number of applications, for example to capture and analyse debates in social networks (Cabrio & Villata, 2013) and to support decision making (Amgoud, Cayrol, & Lagasquie-Schiex, 2008). We describe the use of trained classifiers “in the wild”, analysing two entire, unseen texts. We experiment with various Machine Learning (ML) techniques, i.e. Support Vector Machines (SVMs), Random Forests (RFs), and Long-Short Term Memory (LSTM) networks, also used in (Bosc, Cabrio, & Villata, 2016; Carstens & Toni, 2017; Koreeda, Yanase, Yanai, Sato, & Niwa, 2016) to determine argumentative relations. We train these classifiers on the dataset from (Carstens & Toni, 2015) consisting of pairs of texts annotated as attack, support, or neither attack nor support.

O. Cocarascu (B) · F. Toni Department of Computing, Imperial College London, South Kensington Campus, Huxley Building, London SW7 2AZ, UK e-mail: [email protected] F. Toni e-mail: [email protected] © Springer Nature Switzerland AG 2020 F. H. van Eemeren and B. Garssen (eds.), From Argument Schemes to Argumentative Relations in the Wild, Argumentation Library 35, https://doi.org/10.1007/978-3-030-28367-4_17

269

270

O. Cocarascu and F. Toni

AM has been deployed in various settings, from Twitter (Bosc et al., 2016) to discussion forums (Ghosh, Muresan, Wacholder, Aakhus, & Mitsui, 2014). Our experiments focus on deploying AM to extract BAFs from an excerpt dialogue from the film Twelve Angry Men1 and from a short text, to test whether trained classifiers can achieve comparable outcomes to those of humans. For the first experiment we use the BAFs generated by 21 annotators, enrolled as master Computing students. We did not require the identification of the exact boundaries of arguments within the dialogue, instead we specified these boundaries and asked the annotators to focus on the identification of the relations between arguments. For the second experiment we use the BAFs generated by 19 annotators, again enrolled as master Computing students, from a short text, this time without specifying the boundaries of arguments. In both experiments, we also included in our analysis the BAF that we manually extracted from the texts. For the classifiers we have experimented with, there was no overall winner: LSTMs and RFs performed the closest, for the first and second experiment respectively, to the human annotators. The remainder is organised as follows. We discuss related work in Sect. 17.2 and describe the ML techniques for RbAM and the dataset used for training ML algorithms in Sect. 17.3. We present the first experiment in mining BAFs from the dialogue excerpt from Twelve Angry Men in Sect. 17.4 and the second experiment in mining BAFs from the short text in Sect. 17.5. We conclude and present directions for future work in Sect. 17.6.

17.2 Related Work RbAM, as defined by Carstens and Toni (2015), focuses on the detection of argumentative relations of attack and support and can be seen as a prerequisite for constructing BAFs. BAFs are triples AR, Att, Sup consisting of a set of arguments AR and two binary argumentative relations of attack (Att) and support (Sup) between arguments (Cayrol and Lagasquie-Schiex, 2005). For example, consider the three texts: A1: “We should go to the park.” A2: “It is going to rain.” A3: “We can take an umbrella so we can still go.” These texts can be seen as “arguments” in a BAF with AR = {A1, A2, A3} and relations as shown graphically below (where dashed edges represent attack and standard edges represent support): A1 ← A2 ← A3

1 http://www-sop.inria.fr/NoDE/NoDE-xml.html#12AngryMen.

17 Deploying Machine Learning Classifiers for Argumentative …

271

Whilst most studies that focus on determining links between (components of) arguments assume that the arguments have already been identified, Carstens and Toni (2015) focus on identifying directly the relations between any two texts, irrespective of their stand-alone argumentativeness, being able to identify texts that may appear non-argumentative when analysed in isolation but become argumentative when read in context, as in the earlier simple example. RbAM can be treated as a ML classification task with three classes: support, attack, and neither support nor attack. Several ML techniques have been deployed in the literature for RbAM. Bosc, Cabrio, and Villata (2016) used Long-Short Term Memory (LSTM) (Hochreiter & Schmidhuber, 1997) networks to determine the type of relations between arguments extracted from tweets, whilst Koreeda, Yanase, Yanai, Sato, and Niwa (2016) used a Bidirectional Recurrent Neural Network to classify evidences that support a claim promoting or suppressing a value. Carstens and Toni (2017) used RFs (Breiman, 2001) and SVMs (Boser, Guyon, & Vapnik, 1992) to determine the type of relations between texts extracted from news, whereas Stab and Gurevych (2016) used SVMs to classify support and attack relations in persuasive essays. These ML techniques are also used for other AM tasks we do not consider. For example, Habernal and Gurevych (2016) used LSTMs to determine which argument is more convincing whereas Bowman, Angeli, Potts, and Manning (2015) used stacked LSTMs to determine entailment, neutral and contradiction relations amongst sentence pairs from the SNLI corpus (Bowman, Angeli, Potts, & Manning, 2015). These relations are of a different kind than the ones underlying the construction of BAFs. There are few works on AM “in the wild”. For example, Cerutti, Palmer, Rosenfeld, Snajder, and Toni (2016) manually extract a BAF from a online debate, whereas Salah, Coenen, and Grossi (2013) use sentiment analysis to extract graphs from parliamentary debates where nodes are labelled with the sentiment of the speaker and edges between nodes of same/different polarity are labelled as supporting/opposing.

17.3 Training Classifiers for RbAM 17.3.1 Training Dataset We use a dataset2 covering various topics, with 27% attack relations, 43% support relations, and 30% neither attack nor support relations. The dataset includes, amongst others, a corpus extracted from news, (Carstens & Toni, 2015) and the quotes and replies from the Internet Argument Corpus (IAC) (Walker, Tree, Anand, Abbott, & King, 2012), labelled with three classes (support, attack, neither). This dataset was deemed to be the most appropriate for an experiment “in the wild” as the texts it consists of resemble the ones that can be found in debates and the short 2 https://www.doc.ic.ac.uk/~oc511/ACMToIT2017_dataset.xlsx.

272

O. Cocarascu and F. Toni

Table 17.1 10-fold CV results on training dataset (we report Accuracy, Precision, Recall, and F1) Classifier

A%

P%

R%

F1 %

RF

97.16

97.20

97.16

97.15

SVM

71.34

71.36

71.34

70.99

LSTM

89.53

90.80

87.67

89.07

The variance for F1 using LSTMs is 0.73 on 5 × 10 CV

text that we use in our experiments, unlike persuasive essays (Stab & Gurevych, 2016) and the SNLI corpus (Bowman, Angeli, Potts, & Manning, 2015), which are more structured.

17.3.2 Argument Mining Experiments We experimented with various ML techniques to determine the type of argumentative relations between texts: SVMs, RFs, as well as a deep learning method based on LSTMs. These techniques have been used in previous AM studies as discussed in Sect. 17.2. The deep learning approach uses two parallel LSTMs, where each input text from the pair of arguments considered is fed to a LSTM network as a 100-dimensional GloVe embedding vector (Pennington, Socher, & Manning, 2014). The outputs of the two LSTMs are then concatenated and fed to a fully-connected layer before being fed to a softmax classifier. For the non-deep learning approaches, we use syntactic features such as number of words, the sentiment polarity of each text, and various similarity measures. The results on our training dataset, using 10-fold cross-validation (CV) are given in Table 17.1.

17.4 First Experiment: Mining BAFs from a Dialogue Excerpt We set the task to extract a BAF (by identifying argumentative relations between arguments) from a dialogue excerpt taken from the film Twelve Angry Men, which revolves around the reasoning of twelve men that deliberate on a homicide trial.

17.4.1 Dialogue Excerpt The dialogue excerpt we use is shown below: A0

The boy is not guilty.

17 Deploying Machine Learning Classifiers for Argumentative …

A1 A2 A3 A4 A5 A6

A7

A8 A9

A10

A11 A12 A13 A14 A15 A16

A17

273

I never saw a guiltier man in my life. You sat right in court and heard the same thing I did. The man’s a dangerous killer. He’s 19 years old [too young to be guilty]. Old enough. He knifed his own father, four inches into the chest. An innocent little 19-year old kid. They proved it a dozen different ways. It’s not so easy for me to raise my hand and send a boy off to die without talking about it first. I think the guy’s guilty. You couldn’t change my mind if you talked for a hundred years. This boy’s been kicked around all his life. You know, living in a slum, his mother dead since he was nine. That’s not a very good head start. He’s a tough, angry kid. I think maybe we owe him a few words [before considering him guilty]. We don’t owe him a thing. He got a fair trial. You’re not going to tell us that we’re supposed to believe him, knowing what he is. I’ve lived among ‘em all my life. You can’t believe a word they say. I just think he’s guilty. I thought it was obvious. I mean nobody proved otherwise. Nobody has to prove otherwise. The burden of proof is on the prosecution. The defendant doesn’t have to open his mouth. That’s in the Constitution. The Fifth Amendment. At 10 min after 12 the old man who lived underneath the room where the murder took place heard loud noises in the upstairs apartment. He heard the kid say to his father, “I’m gonna kill you!” A second later he heard a body falling and saw the kid running down the stairs. Then he called the police. They found the father with a knife in his chest. And the coroner fixed the time of death at around midnight. The boy’s story is flimsy. He claimed he was at the movies. That’s a little ridiculous, isn’t it? He couldn’t even remember what pictures he saw. What about the woman across the street? If her testimony [to have seen the killing] don’t prove [that the boy is guilty], then nothing does. [She saw the killing] through the windows of a passing elevated train [so testimony unreliable]. And they proved in court that you can look through the windows of a passing train at night and see what’s happening on the other side. I started to be convinced, with the testimony from those people across the hall. Didn’t they say something about an argument between the father and the boy around 7 o’clock that night? 8 o’clock. They heard the father hit the boy twice and then saw the boy walk angrily out of the house. [It doesn’t prove that the boy killed him].

274

O. Cocarascu and F. Toni

17.4.2 Annotation Results The annotators were 21 master Computing students, with us also extracting a BAF. The students had been lectured on AM, RbAM, and BAFs but were otherwise nonexpert annotators. We instructed the annotators that their annotations should lead to a BAF AR, Att, Sup, with AR = {A0, … , A17}, namely we instructed them that each paragraph in the dialogue should be seen exactly as one argument. We also instructed the annotators that their BAFs should follow the discussion flow in which one argument supports or attacks another argument previously expressed in the dialogue and that the BAF should thus be “minimal” in the sense that each argument (except the first one) has one outgoing relation: ∀a ∈ AR\{A0} ∃b ∈ AR s.t. (a, b) ∈ Att ∪ Sup. We also instructed the annotators that, in the case of an argument supporting and/or attacking more than one argument, to guarantee “minimality”, the relation to the most recent argument only should be included in the BAF. Let < be a total order over AR s.t. Ai < Aj iff i < j, for i, j = 0 … 17. Then, the “minimality” requirement amounts to the following: if a, b1 , b2 ∈ A R and b1 < b2 < a then if (a, b2 ) ∈ Att ∪ Sup then (a, b1 ) ∈ / Att ∪ Sup The reason for preferring relations from more recent arguments is that this allows to limit the number of argumentative relations in BAFs, and thus generate BAFs that can be more easily understood in settings with a vast amount of information such as debates or online comments. Figure 17.1 shows the BAF that we extracted from the excerpt dialogue fulfilling the conditions set for the annotation task. Table 17.2 shows the number of each type of relation identified by all annotators, including our annotation. Note that, whilst given clear instructions for annotation, there were a few annotators who did not follow the “minimality” requirement (i.e. Fig. 17.1 Our manually extracted BAF from the excerpt dialogue (shown graphically)

15s

2a

6s

10a

6s

1a

18a

1s

13a 1s

la

11a

12a

1s

la

10a 1s

4s

A2

A3

A4

A5

A6

A7

A8

A9

A10

A11

A12

A13

A14

A15

A16

A17

4s

4s 1a

1s

4s

3s

1s

la

4s

2a

3s

11a

1a

A1

2s

22a

A2

1s

1s

6a 1s

A3

1s

9s

10a 1s

1s

A4

1s

1s

1s

2s

10a

A5

2a

21a

A6

3s

1s

la

A7

22a

A8

la

1s

la

4a

A9

The relations in bold are the relations that have been identified by at least 6 annotators (our threshold)

20a

A1

A0

A0

3s

2s

4s

18s

A10

la

4s

Al1

2s

A12

4s

21a 1s

A13

18a 1s

A14

Table 17.2 The relations (a for attack and s for support) and the number of times they have been indicated by the annotators

1s

A15

12a 10s

A16

A17

17 Deploying Machine Learning Classifiers for Argumentative … 275

276

O. Cocarascu and F. Toni

(A0, A1), (A2, A4), (A6, A7) violate the requirement). Note also that some relations were identified as both attack and support [e.g. (A17, A16)]. The most controversial pair of arguments is (A17, A16). Indeed, 12 annotators identified this pair as belonging to the attack relation whereas 10 annotators identified it as belonging to support. In all other cases of doubly annotated pairs, one of the annotations is dominant. From now on, we focus only on the relations on which at least 6 annotators agreed. This is because fewer relations than 6 were deemed to be insignificant in evaluating trained classifiers and 6 represents more than a quarter of the number of annotators used in this study. When double annotations occurred, we used majority voting to select the relation type, thus the relation between A17 and A16 was chosen to be attack. All pairs that where doubly annotated can be seen as examples of ambiguous relations: for example, some annotators considered A16 to mean “I started to be convinced that he is not guilty” instead of reading it as “Didn’t they say something about an argument…I started to be convinced that he is guilty”. We used Fleiss’ kappa (Fleiss et al., 1971) and Krippendorff’s alpha (Krippendorff, 2004) to compute the inter-annotator agreement. Both methods gave an agreement of 0.457, showing moderate agreement between the annotators and indicating the difficulty of the task. We also computed the number of relations identified by each annotator compared to the relations identified by at least 6 of the remaining annotators, yielding an average of 60.45% agreement. The relations (and the number of times each relation has been identified) are shown in Fig. 17.2. Fig. 17.2 Argumentative relations and the number of times these have been annotated

17 Deploying Machine Learning Classifiers for Argumentative …

277

17.4.3 Mining Argumentative Relations “in the wild” We report the performance of our trained ML classifiers in determining the type of relations from the ones identified by annotators (in other words, we determine recall). Whilst the ML classifiers identify other relations on top of the ones identified by annotators, we are interested in comparing the argumentative relations identified automatically with the ones identified by annotators. The recall of the methods we experimented with is shown in Fig. 17.3. RFs perform worst, with best recall performance of 22%. This is because RFs fail to learn how to classify attack relations. This can also be seen in the graph as all relations that have been identified by at least 20 annotators are attacks and RF fails to classify these correctly. On the other hand, SVMs achieve good results classifying all relations as attack, which is the dominant type of relation in our BAF. The LSTM-based method is the one that performs the best, being able to classify both attack and support relations with recall above 80%. Our deep network is also able to classify argumentative relations that had not been identified by annotators (as they would have given rise to a non-“minimal” BAF). For example, our network shows that A16 supports A10, and that A14 supports A6. Whether identifying all possible relations between all arguments is better than extracting a “minimal” BAF is left for future work. Fig. 17.3 Recall of ML classifiers we used to determine argumentative relations. Threshold t represents the number of times the relations have been identified by at least t annotators

278

O. Cocarascu and F. Toni

17.5 Mining Argumentative Relations from a Short Text Our second experiment focus on testing whether ML classifiers also perform similarly to humans when deployed on short texts. We asked 19 annotators to manually extract a BAF, with no information about the boundaries of arguments, from the following text which presents reasons for and against a person (Mary) renting some flat3 : B0 B1 B2 B3 B4 B5 B6 B7 B8 B9 B10 B11

Mary should rent the flat. The flat is near Mary’s work. The flat is located in a well illuminated and safe area. The flat is in a quiet area, because most of the neighbours are retired. The flat is tiny. Despite the size, spaces are well distributed within the flat. The flat shows signs of humidity problems. There are rumours that a nightclub will open in the area. Humidity problems take time to resolve. Local legislation forbids the opening of a nightclub in the area. The building’s insurance covers fixing any problems promptly. A renowned architect designed the building to optimise space distribution.

Figure 17.4 gives the BAF that we manually extracted from this text. Without being given specific instructions about the boundaries of arguments, several annotators identified arguments at sentence level, whereas others split the sentences into multiple “arguments”. We report the “merged” version of arguments in Sect. 17.5.1 and the “unmerged” version of arguments in Sect. 17.5.2.

17.5.1 Merged Arguments Table 17.3 shows the number of each type of relation identified by annotators (support/attack) and Fig. 17.5 shows the relations (and the number of times each relation Fig. 17.4 Our manually extracted BAF from the flat example

3 The

sentence labelling is given here for ease of reference, and was not given to the annotators, who instead were presented with a monolithic, but short, text.

19s

19s

19a

B2

B3

B4

2a

19a 1s

B4

1a 18s

B5

1s

B6

1s

B11

9a 1s

The relations in bold are the relations that have been identified by at least 6 annotators (our threshold)

2s

B10

1a

1s

8a 1s

B3

B9

B2

3a 14s

9a

B7

la

1a

1s

1s

B1

B8

19a

B6

B5

20s

B1

B0

B0

16a

1a

B7

8a

1s

B8

B9

B10

B11

Table 17.3 The relations (a for attack and s for support) and the number of times they have been indicated by the annotators for the “merged” version

17 Deploying Machine Learning Classifiers for Argumentative … 279

280

O. Cocarascu and F. Toni

Fig. 17.5 Argumentative relations and the number of times these have been annotated for the “merged” version

has been identified). The inter-annotator agreement using Fleiss’ 578 kappa and Krippendorff’s alpha is 0.655, showing substantial agreement. The number of relations identified by each annotator compared to the relations identified by at least 6 of the remaining annotators yields an average of 79.11% agreement. The recall of our ML classifiers is shown in Fig. 17.6. RFs perform the best, identifying 77% of the relations labelled by at least 14 annotators, being able to identify both support and attack relations, unlike LSTM and SVM which classify all relations as attack. We believe this is because the length of the arguments in this experiment was shorter than those in our training dataset. Fig. 17.6 Recall of ML classifiers we used to determine argumentative relations in the “merged” version of the short text

17 Deploying Machine Learning Classifiers for Argumentative …

281

17.5.2 Unmerged Arguments As explained in Sect. 17.5, some annotators split some sentences in different arguments. This applies to B2 and B3 as follows: B2 The flat is located in a well illuminated and safe area. is identified in three forms: B2 The flat is located in a well illuminated and safe area. B12 The flat is located in a well illuminated area. B13 The flat is located in a safe area. Whereas: B3 The flat is in a quiet area, because most of the neighbours are retired. is identified in three forms: B3 The flat is in a quiet area, because most of the neighbours are retired. B14 The flat is in a quiet area. B15 Most of the neighbours are retired. Table 17.4 gives the annotations when we consider split and unsplit arguments. The relations (and the number of times each relation has been identified) are shown in Fig. 17.7. The inter-annotator agreement using Fleiss’ kappa is 0.564 and Krippendorff’s alpha is 0.565, showing moderate agreement. The number of relations identified by each annotator compared to the relations identified by at least 6 of the remaining annotators yields an average of 74.26% agreement. The recall of our ML classifiers is shown in Fig. 17.8. Again, RFs perform the best, identifying 75% of the relations that have been identified by at least 14 annotators being able to identify both support and attack relations, unlike LSTM and SVM which classify all relations as attack. RFs identify the support relation (B14, B0) and (B15, B14) as well as the relation (B3, B0) related to argument B3 which has been split into two arguments by the annotators. The relations from arguments B12 and B13 are not considered as these had been identified by only 2 annotators, which is smaller than our threshold of at least 6 annotators.

17.6 Conclusion We described two experiments “in the wild” on using different trained classifiers, SVMs (Boser et al., 1992), RFs (Breiman, 2001) and LSTM networks (Hochreiter and Schmidhuber, 1997), to determine argumentative relations of support and attack in text. We asked human annotators to identify the argumentative relations from an excerpt dialogue taken from the movie Twelve Angry Men and from a short text, with the

17s

7s

19a

B3

B4

2a

19a 1s

B4

la 18s

B5

1s

B6

1s

2s

2s

12s

B11

B12

B13

B14

9a 1s

16a

1a

B7

8a

1s

B8

B9

The relations in bold are the relations that have been identified by at least 6 annotators (our threshold)

B15

2s

B10

1a

1s

3a 1s

B3

B9

1a

B2

3a 14s

9a

B7

1a

1s

1s

B1

B8

19a

B6

B5

20s

B2

B0

B1

B0

B10

B11

B12

Table 17.4 The relations (a for attacks and s for supports) and the number of times they have been indicated by the annotators B13

12s

5a

B14

B15

282 O. Cocarascu and F. Toni

17 Deploying Machine Learning Classifiers for Argumentative …

283

Fig. 17.7 Argumentative relations and the number of times these have been annotated for the “unmerged” version

Fig. 17.8 Recall of ML classifiers we used to determine argumentative relations in the “unmerged” version of the short text

aim of extracting Bipolar Argumentation Frameworks (Cayrol & Lagasquie-Schiex, 2005) from each. We compared the argumentative relations extracted using trained classifiers with the ones extracted by the annotators and showed that there is no overall winner as to which ML classifier is better suited for this task: LSTMs and RFs performed the best, for the first and second experiment respectively, suggesting that different techniques might be better suited for different types of texts. Annotating argumentative relations is a cumbersome task, difficult for human annotators, even for trained ones (Peldszus & Stede, 2013). Future work includes testing the classifiers we considered and others on other texts. We aim at developing classifiers that achieve comparable results to human annotators in a wide range of settings. These classifiers could then be used to create more corpora, highly needed for experimentation. We have assumed that arguments are represented by at least one sentence. Future work includes identifying the arguments relevant for the issue considered (Wachsmuth, Stein, & Ajjour, 2017).

284

O. Cocarascu and F. Toni

References Amgoud, L., Cayrol, C., Lagasquie-Schiex, M. C., & Livet, P. (2008). On bipolarity in argumentation frameworks. International Journal of Intelligent Systems, 23(10), 1062–1093. Bosc, T., Cabrio, E., & Villata, S. (2016). Tweeties squabbling: Positive and negative results in applying argument mining on social media. In Computational models of argument COMMA (pp. 21–32). Boser, B. E., Guyon, I. M., & Vapnik, V. N., (1992). A training algorithm for optimal margin classifiers. In Proceedings of the Fifth Annual Workshop on Computational Learning Theory. COLT ‘92, (pp. 144–152). Bowman, S. R., Angeli, G., Potts, C., & Manning, C. D. (2015). A large annotated corpus for learning natural language inference. In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing (EMNLP). Association for Computational Linguistics. Breiman, L. (2001). Random forests. Machine Learning, 45(1), 5–32. Cabrio, E., & Villata, S. (2013). A natural language bipolar argumentation approach to support users in online debate interac-tions. Argument & Computation, 4(3), 209–230. Carstens, L., & Toni, F. (2015). Towards relation based argumentation mining. In Proceedings of the 2nd Workshop on Argumentation Mining (pp. 29–34). Carstens, L., & Toni, F. (2017). Using argumentation to improve classification in natural language problems. ACM Transactions on Internet Technology, 17(3), 30, 1–30, 23. Cayrol, C., & Lagasquie-Schiex, M.-C. (2005). On the acceptability of arguments in bipolar argumentation frameworks. In Symbolic and quantitative approaches to reasoning with uncertainty: 8th European conference (pp. 378–389). Cerutti, F., Palmer, A., Rosenfeld, A., Snajder, J., & Toni, F. (2016). A pilot study in using argumentation frameworks for online debates. In Proceedings of the First International Workshop on Systems and Algorithms for Formal Argumentation (SAFA) (Vol. 1672, pp. 63–74). CEURWS.org. Fleiss, J. L., et al. (1971). Measuring nominal scale agreement among many raters. Psychological Bulletin, 76(5), 378–382. Ghosh, D., Muresan, S., Wacholder, N., Aakhus, M., & Mitsui, M. (2014). Analyzing argumentative discourse units in online interactions. In Proceedings of the First Workshop on Argument Mining (pp. 39–48). Habernal, I., & Gurevych, I. (2016). What makes a convincing argument? empirical analysis and detecting attributes of convincingness in web argumentation. In Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing (pp. 1214–1223), EMNLP. Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural Computation, 9(8), 1735–1780. Koreeda, Y., Yanase, T., Yanai, K., Sato, M., & Niwa, Y. (2016). Neural attention model for classification of sentences that support promoting/suppressing relationship. In Proceedings of the Third Workshop on Argument Mining. Krippendorff, K. (2004). Content analysis: An introduction to its methodology (second edition). Sage Publications. Lippi, Marco, & Torroni, Paolo. (2016). Argumentation mining: State of the art and emerging trends. ACM Transactions on Internet Technology, 16(2), 10. Peldszus, A., & Stede, M. (2013). Ranking the annotators: An agreement study on argumentation structure. In: Proceedings of the 7th Linguistic Annotation Workshop and Interoperability with Discourse (pp. 196–204). Pennington, J., Socher, R., & Manning, C. D. (2014). Glove: Global vectors for word representation. In Empirical methods in natural language processing (EMNLP) (pp. 1532–1543). Salah, Z., Coenen, F., & Grossi, D. (2013). Extracting debate graphs from parliamentary transcripts: A study directed at uk house of commons debates. In Proceedings of the Fourteenth International Conference on Artificial Intelligence and Law (pp. 121–130).

17 Deploying Machine Learning Classifiers for Argumentative …

285

Stab, C., & Gurevych, I. (2016). Parsing argumentation structures in persuasive essays. CoRR abs/1604.07370. Wachsmuth, H., Stein, B., & Ajjour, Y. (2017). “PageRank” for argument relevance. In Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics (EACL 17) (pp. 1116–1126). Walker, M. A., Tree, J. E. F., Anand, P., Abbott, R., & King, J. (2012). A corpus for research on deliberation and debate. In Proceedings of the Eighth International Conference on Language Resources and Evaluation, LREC (pp. 812–817).

Index

A Abduction, 7, 205–209, 213, 214, 216, 218–220 Abductive trigger, 206, 218, 220 Accessibility, 46–50, 59, 60 Analogy argumentation, 6, 11 Analytic dilemma, 4, 84, 90, 91, 94, 95 Appeal to authority, 20 Apprehensive use, 187 Arguing for question, 6 Argumentative relation, 1, 8, 9, 269, 270, 272, 274, 276, 277, 280, 281, 283 Argument mining, 8, 9, 269, 272 Argument scheme / argumentative scheme, 1, 2, 8, 9, 11–15, 17–19, 21, 25, 28, 32, 39, 44, 125, 194, 199, 201, 202, 269, 270, 272, 274, 276–278, 280, 281, 283 Argument strength, 143, 147, 149, 262–266 Auxiliary condition, 2, 25, 26, 33, 38 Auxiliary expert, 2, 26, 27, 29–39

B Backgrounding, 3, 42, 48–50 Backing, 5, 131–133, 136–138, 140, 141, 160, 161 Balance-of-considerations argument, 256 Basic argument scheme, 28, 32, 39 Burden of proof, 3, 4, 7, 59, 68, 79, 81–85, 90, 91, 93–95, 169, 174, 201, 214, 215, 220, 273

C Causal argumentation, 6, 15–19, 193

Character of arguer, 98 Cogency, 4, 5, 97, 137, 144, 147, 150 Cognitive process, 3, 42, 46, 48 Community, 1, 3–5, 27, 31, 35, 36, 97, 100, 102, 105, 107–117, 121, 122, 124, 125, 127 Conduction, 8, 239–244, 246–250, 252, 255–267 Conductive argument, 7, 8, 223, 226–228, 235, 236, 239–253, 255–266 Conversational implicature, 209, 210, 212, 215 Conversational maxim, 51, 104, 209, 212 Counter-analogy, 152 Counter-argument, 4–6, 57, 71, 83, 249 Counter-consideration, 3, 7, 8, 80, 81, 225–231, 234, 236, 241–253, 257–261 Coverage sufficiency, 36, 38 Critical discussion, 2, 3, 12–15, 20, 49, 64, 71, 73–75, 170, 197, 200, 202, 205, 206, 208, 213–216, 218, 220, 252 Critical rationalism, 223

D Decision-making, 9, 125, 207, 223–225, 269 Decisive objection, 7, 225–236 Defeasible argument, 147, 247, 258, 259, 262, 264–266 Deliberation, 112, 224, 225, 229, 231, 232, 235, 261 Dialectical adversariality, 104 Dialectical profile, 16, 20 Dialectical table, 85

© Springer Nature Switzerland AG 2020 F. H. van Eemeren and B. Garssen (eds.), From Argument Schemes to Argumentative Relations in the Wild, Argumentation Library 35, https://doi.org/10.1007/978-3-030-28367-4

287

288 Dialectics, 5, 63, 64, 81, 90, 95, 125, 149, 159, 240 Dialogical sequence, 3, 4, 80, 81, 84, 85, 93–95 Disclosing irrelevant distinction, 4 Dismissal, 5, 150, 151 Distraction, 4, 52, 108, 116–118, 124–127

E Epistemic modality, 192 Epistemic strength, 2, 42, 46–50, 52, 60 Erotetic inference, 170, 171, 179–181 Eudaimonia, 98, 100, 102 Evocation, 172, 173 Evolution, 110, 111, 114–116, 181

F Fallacy, 1–3, 6, 39, 41–45, 47–52, 54, 57, 60, 63–75, 81, 126, 127, 151, 192, 194, 195, 197, 199–202, 206, 212, 214–216, 219, 220 Foregrounding, 3, 42, 48–50

Index M Machine learning, 9, 269 Maxim, 51, 104, 200, 209, 210, 212, 213 Mechanism of question generation, 79, 81

N Narrative argument, 108, 125 Negative consideration, 8, 157, 223, 239, 240, 256, 258–267 Non-basic argument scheme, 28 Non-testimony, 35, 36 Nudging, 4, 108, 118, 119

O Objection, 2, 5–7, 32, 49, 65, 70, 71, 81, 82, 117, 122, 150–152, 157, 159–164, 174, 181, 223, 225–232, 234–236, 261, 265, 266 Objectivity, 100 On-balance premise, 8, 240–242, 245, 246, 248–253, 262 Open-ended question, 173, 176, 180, 182 Overt, 7, 49, 199–201

G Genetic fallacy, 2, 3, 64, 66–75

H Human flourishing, 3, 4, 97, 98 Hyperbole, 199, 200

I Immunization, 6, 186, 197, 198, 201, 202 Inferential erotetic logic, 6, 170, 182 Institution, 5, 55, 119, 131, 133, 145–147 Intellectual flourishing, 101–105 Interpretation, 7, 26–28, 30, 31, 33, 38, 39, 47, 51, 52, 54, 56, 58, 60, 74, 80, 85, 90, 94, 133, 139, 169–171, 181, 205, 206, 208–211, 213, 215, 218–220, 244, 246, 257, 260 Interrogative act, 168–170, 173–176, 178 Interrogative sentence, 167–171, 177, 178 Irony, 200, 205, 209–213, 215, 216, 218, 220

L Layman’s paradox, 34, 36–38 Logical reconstruction, 8, 240

P Persuasion, 4, 5, 47, 50, 51, 53, 100, 101, 108, 110, 116–121, 123–126, 240 Philosophical argumentation, 3, 4, 66, 75, 80, 81, 83, 95 Philosophical argumentative move, 3, 4, 83 Philosophical dialogue, 3, 65 Plausibility, 47, 138, 144, 147, 161, 243, 244, 252 Practical proposal, 225 Practical reasoning, 102, 229 Praeteritio, 201 Pragma-dialectics, 7, 15, 125, 206, 209, 213, 214, 216, 218–220 Pragmatic argumentation, 6, 17–19, 193, 194, 202 Pragmatics, 2, 4, 6, 13–15, 17–19, 42, 50–52, 68, 84, 92, 94, 95, 132, 156, 163, 168, 175–177, 182, 193, 194, 202, 209, 210, 212, 216, 218, 219, 244, 245, 250, 252 Pragmatic self-refutation, 4, 84, 92, 94, 95 Pro/con argument, 5, 7, 120, 149, 150, 164, 223, 225–227, 229, 230, 235, 256 Profile of dialogue, 81, 159

Index R Rebuttal, 5, 6, 81, 82, 85, 88, 89, 137–141, 144–147, 150, 152, 153, 157–164, 229, 230, 242, 263 Rebuttal tolerance, 145, 147 Reductio ad Absurdum, 192, 195, 196, 199, 202 Refutation, 5, 6, 12, 42, 48, 57, 58, 82, 117, 150, 152, 155–164, 251 Relevance theory, 44, 47, 52, 210 Rhetorical device, 205, 215 Rhetorical effectiveness, 3, 41, 42, 44, 45, 47, 50, 52, 53, 60

S Scheme integrity, 28, 39 Slippery slope, 36, 192–196, 199 Sociality, 110, 111, 113, 126 Speech act, 134, 194, 209, 213–216, 243 Strategic maneuvering / strategic manoeuvring, 2, 3, 6, 22, 51, 64, 68, 70–74, 185, 186, 192, 195, 196, 201, 202, 206, 215, 216, 218–220, 261 Straw man, 51, 57–59, 74 Strength of argument, 8, 37, 141, 143, 144, 147, 149, 157, 158, 259, 262–266

289 Symptomatic argumentation, 15, 16, 19

T Third pattern of conduction, 157, 239, 240, 256 Toulmin model, 117, 125, 242, 243, 252

U Untruthfulness, 200

V Valid inference, 174, 175, 177, 178 Virtue epistemology, 4, 97, 98, 101 Virtue theory of argumentation, 4, 97–101, 103, 105 Visual argument, 108, 122, 123, 125

W Warrant, 5, 8, 79–82, 85, 117, 121, 131–147, 152–154, 160, 161, 163, 187, 225, 240, 242–245, 252 Weighing, 83, 156–158, 163, 226, 240–242, 246, 248, 251, 252, 260, 261, 265