Modality in Argumentation PDF

Argumentation Library Andrea Rocci Modality in Argumentation A Semantic Investigation of the Role of Modalities in the Structure of Arguments with an Application to Italian Modal Expressions Argumentation Library Volume 29 Series Editor Frans H. van Eemeren, University of Amsterdam, The Netherlands Editorial Board Bart Garssen, University of Amsterdam, The Netherlands Scott Jacobs, University of Illinois at Urbana-Campaign, USA Erik C.W. Krabbe, University of Groningen, The Netherlands John Woods, University of British Columbia, Canada Since 1986 Springer, formerly Kluwer Academic Publishers, publishes the interna- tional interdisciplinary journal Argumentation. This journal is a medium for distrib- uting 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 theoreti- cal 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, contribu- tions are invited from argumentation theorists with a background in informal or formal logic, modern or classical rhetoric, and discourse analysis or speech com- munication. 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 argu- mentation 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. More information about this series at http://www.springer.com/series/5642 Andrea Rocci Modality in Argumentation A Semantic Investigation of the Role of Modalities in the Structure of Arguments with an Application to Italian Modal Expressions Andrea Rocci Istituto di Argomentazione, Linguistica e Semiotica Università della Svizzera italiana University of Lugano Lugano, Ticino, Switzerland ISSN 1566-7650 ISSN 2215-1907 (electronic) Argumentation Library ISBN 978-94-024-1061-7 ISBN 978-94-024-1063-1 (eBook) DOI 10.1007/978-94-024-1063-1 Library of Congress Control Number: 2016963798 © Springer Science+Business Media B.V. 2017 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, express 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. Printed on acid-free paper This Springer imprint is published by Springer Nature The registered company is Springer Science+Business Media B.V. The registered company address is: Van Godewijckstraat 30, 3311 GX Dordrecht, The Netherlands Acknowledgments This book would not have been started without the instigation of Eddo Rigotti, my teacher and then colleague and friend. He introduced me to linguistics, to semantics, and to the treasures of ancient and medieval thinking about language, logic, argu- ment and rhetoric. And he bluntly told me that this book had to be written. He instigated so many good things in my life that it will be hard to thank him enough in this life. I’m happy to be able to do that for this book. This book would never have been finished without the encouragement, wisdom, and friendship of Frans van Eemeren. Without Frans I would not be doing argumentation theory in the first place – and the field itself would be very different without his monumental contribution. It took a long time to write this book, and there were times when I was close to giving up on the project. Frans’ belief in the project was a key factor in bringing it to completion. My Lugano colleague Johanna Miecznikowski, a true linguist and one of the nic- est persons I have ever met, is also to thank. She was a driving force in the research projects from which this book is born and coauthor of important papers. But I thank her even more for the many pleasant conversations on abstruse issues of modality through these years. Our former PhD student Elena Musi came to share these con- versations and, after moving to the USA, she is bringing forward the research pro- gram at the heart of this book in ways I had not foreseen. Besides being an energetic researcher and coauthor, Elena took the time to help me with the task of bringing this ponderous manuscript, evolved through many drafts, to publishable form. When Elena moved to the USA, Chiara Pollaroli, another former PhD student and research collaborator, took up this task and was at my side in the latest editing stages. I am enormously grateful for that. This research would not have been possible without the generous support of the Swiss National Science Foundation (SNF) with the grants 100012-120740/1 and 141350. The University of Lugano – also known by its official name of Università della Svizzera italiana (USI) – is a lively interdisciplinary working environment where, thanks to the initiative of Eddo Rigotti, it was possible to develop argumentation as a field of research within the Faculty of Communication Sciences. I am grateful to v vi Acknowledgments several Lugano colleagues and former colleagues, but particularly to Sara Greco and Rudi Palmieri, with whom I have shared the development of the argumentative adventure started by Eddo. To my wife Alessandra and to my children Giovanni, Stefano, and Maria goes my deepest gratitude and love, not because they endured the writing of this long book with its troublesome impact on family life (which they did), but rather because they are the warp and weft of a life where even such an endeavor can be meaningful. ∗∗∗ Being the final report of about a decade of research, this book also incorporates and updates results that have seen publication elsewhere. In particular, the initial sections of Chap. 3 develop ideas initially presented in: –– Rocci, Andrea. 2008. Modality and its conversational backgrounds in the recon- struction of argumentation. Argumentation 22: 165–189. Similarly, the analyses of the Italian modal verbs proposed in Chap. 6 are expanded and revised versions of those presented in previous publications. In par- ticular, the chapter reproduces passages from: –– Rocci, Andrea. 2012. Modality and argumentative discourse relations: a study of the Italian necessity modal dovere. Journal of Pragmatics 44(15): 2129–2149. –– Rocci, Andrea. 2013. Modal conversational backgrounds and evidential bases in predictions: The view from the Italian modals. In Time: language, cognition & Reality, eds. Kasia M. Jaszczolt, and Louis de Saussure, 128–153. Oxford: Oxford University Press. Contents 1 Introduction................................................................................................ 1 1.1 Argumentation and Modality: Working Definitions........................... 2 1.2 Modality and the Essential Structure of Arguments: A First Reconnaissance................................................................................... 4 1.3 Aims of the Study............................................................................... 8 1.4 A Twofold Research Strategy............................................................. 12 1.5 A Focus on Italian Linguistic Structures and on Contextualized Discourse Data...................................................... 16 1.6 What This Book Is Not About............................................................ 17 1.6.1 This Book Is Not About Modal Logic.................................... 18 1.6.2 This Book Is Not About Non-demonstrative and Non-deductive Reasoning................................................ 23 1.7 Structure of the Volume and Reading Guide...................................... 26 References.................................................................................................... 29 2 Meaning and Argumentation.................................................................... 33 2.1 The Viewpoint of Argumentation Theory........................................... 33 2.1.1 Argumentation Theory Is Normative...................................... 34 2.1.2 Dialectic and Normative Pragmatics...................................... 36 2.1.3 The Inferential Dimension...................................................... 42 2.2 Argumentative Analysis...................................................................... 44 2.2.1 Analysis as a Precondition of Critical Evaluation.................. 44 2.2.2 What an Argumentative Analysis Needs to Capture............... 45 2.2.3 Using AMT to Reconstruct the Inferential Configuration of Enthymematic Arguments........................... 48 2.3 The Place of Pragmatic and Semantic Considerations in the Analysis of Argumentative Discourse...................................... 62 2.3.1 The Interplay of Meaning Analysis and Argument Criticism: An Illustration........................................................ 62 2.3.2 Pragmatics and Semantics in Argumentative Reconstruction........................................................................ 67 vii viii Contents 2.3.3 Semantic Types of the Propositions Functioning as Standpoints and as Premises............................................... 73 2.3.4 Congruity Theory.................................................................... 82 2.4 The Sociopragmatic Dimension: Argumentation in Context.............. 95 References.................................................................................................... 98 3 Three Views of Modality in Toulmin........................................................ 105 3.1 Toulmin and Modality Beyond the Toulmin Model........................... 105 3.1.1 The Place of Modality in Stephen Toulmin’s The Uses of Argument............................................................. 106 3.1.2 The Uses of Toulmin’s Views in This Book........................... 107 3.2 The First View: Modal Forces, Criteria and Fields of Argument........................................................................................ 108 3.2.1 Modal Forces and the Phases of an Argument....................... 108 3.2.2 From Modal Criteria to Fields of Argument and Logical Types................................................................... 112 3.2.3 Cannot: An Early Analysis of Modal Context-Dependency.............................................................. 117 3.2.4 First Interlude: Salvaging the Pragmatic Account of the Modals Through the Polyphony and Delocutivity Route........................................................... 124 3.2.5 On the Irrelevance of Logical Modalities............................... 125 3.3 The Second View: Probability and Speech Acts................................. 131 3.3.1 Probably and Guarded Commitment...................................... 131 3.3.2 Ennis vs. Freeman on Toulmin’s Probably............................. 138 3.3.3 Second Interlude: Modality as Detached from the Propositional Content............................................... 146 3.4 The Third View: Modal Qualifiers and the Layout of Arguments...................................................................................... 157 3.4.1 The Toulmin Model................................................................ 157 3.4.2 Third Interlude: The Modal Qualifier and the Place of Modality in the Logical Tradition...................................... 161 3.4.3 Modal Qualifiers and Warrants............................................... 171 3.4.4 Modal Qualifiers and Rebuttals.............................................. 173 3.5 Toulmin’s Three Takes on Modality and Their Legacy in Contemporary Argumentation Theory............................................ 181 3.5.1 Addressing the Toulminian Research Questions.................... 181 3.5.2 The Narrow View: Argumentatively Relevant Modals Are Epistemic Qualifiers............................................ 182 3.5.3 The Epistemic Qualification of Standpoints in Pragma-Dialectics............................................................... 185 3.5.4 Epistemic Qualifiers as Presentational Devices...................... 187 3.5.5 Beyond Epistemic Scales........................................................ 191 References.................................................................................................... 192 Contents ix 4 Relative Modality and Argumentation.................................................... 197 4.1 Modal Semantics and Its Argumentative Implications....................... 197 4.2 The Theory of Relative Modality: From Possible Worlds to Discourse Structure............................................................ 199 4.2.1 Alternatives and Domains of Quantification.......................... 199 4.2.2 Modals Are Relational............................................................ 202 4.2.3 Modals Are Context-Dependent Expressions Whose Valence Is Saturated in Context.................................. 203 4.2.4 Modal Restrictors Are ad hoc and Vague............................... 205 4.2.5 Interlude: Context Dependency in Language......................... 208 4.2.6 The Analysis of Restrictors: From Frame Semantics to Premise Semantics............................................ 211 4.2.7 Possibility and Necessity as Relative to a Conversational Background............................................. 215 4.2.8 Relative Modality and Conditionals....................................... 222 4.2.9 Relative Modality, Anaphora and Discourse Relations.......... 229 4.3 Argumentative Implications of Relative Modality............................. 233 4.3.1 Reconstructing Modalities as Part of the Propositional Content of Standpoints........................................................... 233 4.3.2 Quantifiers as Argumentative Indicators: Revisiting Snoeck Henkemans’ Analysis............................... 234 4.3.3 Propositional Modals as Argumentative Indicators of Argumentation Structure and Beyond................................ 240 4.3.4 Putting Order into the Conversational Background................ 249 4.3.5 Divergence, Vagueness and Shifts in Conversational Backgrounds........................................................................... 254 4.3.6 Modals as Indicators of Argumentative Discourse Relations................................................................ 258 References.................................................................................................... 271 5 Types of Conversational Backgrounds and Arguments......................... 275 5.1 Introduction......................................................................................... 275 5.2 Towards a Typology of Conversational Backgrounds........................ 277 5.2.1 Types of Linguistic Manifestations of the Conversational Background.......................................... 277 5.2.2 Criteria for a Typology of Conversational Backgrounds........ 280 5.2.3 Linguistic-Semantic Classifications of Modal Flavors........... 284 5.2.4 A classification Balancing Linguistic and Logical Criteria................................................................ 291 5.3 Types of Conversational Backgrounds and Arguments...................... 308 5.3.1 Alethic Conversational Backgrounds and Arguments............ 308 5.3.2 Deontic Conversational Backgrounds and Arguments........... 320 5.3.3 The Anankastic Modalities: Necessary Conditions and Telos................................................................................. 346 5.3.4 Epistemic Conversational Backgrounds in Arguments.......... 351 References.................................................................................................... 365 x Contents 6 Case Studies of Italian Modal Constructions in Context....................... 371 6.1 Introduction to the Case Studies......................................................... 371 6.1.1 Taking Advantage of Native-Speaker Competence................ 372 6.1.2 A Corpus of Italian Financial News Articles.......................... 373 6.1.3 An in-Depth Look at Specific Necessity Constructions in Italian.................................................................................. 377 6.2 Argumentation and the Genre of Financial News Articles................. 387 6.2.1 The Financial Markets: Interaction Field and Discourse System............................................................. 387 6.2.2 Prediction and the Discourse Organisation of Financial News................................................................... 389 6.2.3 Modality and Prediction in Financial News........................... 400 6.3 DEVEE vs. DOVREBBEE: Two Kinds of Apparently Epistemic Readings for the Necessity Modal Verb............................. 407 6.3.1 Constraints on Future Reference, Causality and Argumentative Connective Predicates............................. 410 6.3.2 Hypotheses on French DEVRAITE and English SHOULDE........................................................... 416 6.4 The Argumentative Relevance of Non-epistemic Modal Verbs in Predictions................................................................. 418 6.4.1 Futurity as Modality and the Propositional Content of Predictive Standpoints........................................................ 418 6.4.2 Evidential Implications of Future-Oriented Alethic and Deontic Modals................................................................ 424 6.4.3 Economic Causality: Necessity and Impossibility................. 429 6.4.4 Quantificational Readings of the Possibility Modal............... 431 6.4.5 Economic Circumstances and Agent’s Goals......................... 432 6.4.6 Deontic Readings of potere.................................................... 433 6.4.7 Deontic Readings of dovere: Schedules and Plans................. 434 6.5 Semantic Analysis and Argumentative Functioning of DEVEE and DOVREBBEE............................................................... 437 6.5.1 Refining the Hypothesis in the Light of the Non-epistemic Readings of the Modals.................................. 437 6.5.2 Dovrebbe in the Corpus.......................................................... 440 6.5.3 Semantic Analysis................................................................... 440 6.5.4 A Corpus Study of the Conversational Backgrounds of DOVREBBEE in Financial News........................................ 451 6.5.5 The Endoxical Nature of the Normal Conditions C in DOVREBBEE..................................................................... 456 6.5.6 The Inferential Function of C in DOVREBBEE..................... 458 References.................................................................................................... 460 Contents xi 7 Conclusion.................................................................................................. 467 7.1 Tacking Stock..................................................................................... 467 7.2 The Three Questions........................................................................... 468 7.2.1 (Q1) How Are Modals Similar to Arguments?....................... 468 7.2.2 (Q2) Is There a Special Place for Modality in the Structure of Arguments?............................................... 470 7.2.3 (Q3) What Does the Use of Modal Expressions in an Ordinary Discourse Tell Us About the Arguments Being Put Forth in the Discourse?.......................................... 475 7.3 Developments..................................................................................... 479 References.................................................................................................... 481 Index.................................................................................................................. 483 Chapter 1 Introduction Abstract This short chapter introduces the notions of modality and argumentation and lays down the central research questions of the book, which revolve around modality in relation to the essential structure of arguments. The Introduction also serves as a disclaimer explaining how the book is neither a contribution to modal logic nor to the theory of non-monotonic, presumptive reasoning. The final section provides a reading guide previewing the following chapters. τά γὰρ πολλὰ περὶ ὧν αἱ κρίσεις καὶ αἱ σκέπσεις, ἐνδέχεται καὶ ἄλλως ἔχειν “for most of the things which we judge and examine can be other than they are” (Aristotle, Rhetoric, I, 1357a, transl. by J.H. Freese, LOEB1) These terms – ‘possible’, ‘necessary’ and the like – are best understood, I shall argue, by examining the functions they have when we come to set out our arguments. (Toulmin 2003 : 17) In fact, one of the main virtues of a premise semantics for modality is that it links the semantics of modals to general principles of rational inquiry that apply whenever we reason from a set of premises. (Kratzer 2012 : 10) This book examines modality – the semantic category encompassing notions of possibility and necessity – looking at its role in argumentation, that is in discourse where arguments (or reasons) are given in support of a standpoint (or conclusion).2 One of the founders of modern argumentation theory, Stephen E. Toulmin, quoted above in epigraph, chose to open his inquiry into the “working logic”3 of 1 See Aristotle (1926) in the references. Throughout the book, for the translation of passages from the works of Aristotle, I will normally refer to the Revised Oxford Translation edited by Jonathan Barnes (see Aristotle 2014). On a few occasions, like the quote in epigraph to this Introduction, I will refer to another translation that appears more transparent for the purposes at hand. 2 This book represents the (belated) “final report” of the research project Modality in Argumentation. A semantic-argumentative study of predictions in Italian economic-financial newspapers. The project was generously supported by the Swiss National Science Foundation (Grant: 100012- 120740/1) from September 2008 to December 2011. Andrea Rocci was the principal investigator, Eddo Rigotti co-applicant. Johanna Miecznikowski was senior researcher and Gergana Zlatkova participated as Ph.D. student. 3 The expression “working logic” is used in the title of the IV essay of the book, where it is opposed to the “idealised logic” informed by the standard of logical necessity/ deductive validity (cf. Toulmin 2003 : 135). © Springer Science+Business Media B.V. 2017 1 A. Rocci, Modality in Argumentation, Argumentation Library 29, DOI 10.1007/978-94-024-1063-1_1 2 1 Introduction real-life “fields of argument” with a discussion of the meaning of modal expres- sions. If he chose to open The Uses of Argument with such a discussion and then continued to refer to modality throughout the book, it is certainly because he believed that modal notions are particularly revealing of the nature of ordinary argu- ments. In particular, in the first essay the discussion of the meaning of the modals was instrumental to introduce the notion of different “fields” from which real world arguments draw the criteria that justify their force. It is according to these field- dependent criteria, Toulmin insists, that real arguments should be evaluated and not according to the logician’s ideal of deductive validity. Yet, it is interesting to observe that Toulmin, in the quote above, does not say that one can better understand arguments by examining the modals occurring in them. He takes the converse perspective, saying that the modal expressions “are best understood […] by examining the functions they have when we come to set out our arguments”. What follows this quote is an insightful, yet problematic, semantic analysis of the English modal expression cannot, which I will dissect in Chap. 3, § 3.2.3. Similarly, the second essay of Toulmin’s book reads equally as an analysis of the semantics and pragmatics of the adverb probably and as a discussion of a notion of probability relevant for ordinary arguments. The present investigation also takes both perspectives. On the one hand, it exam- ines what modals tell us about arguments, and reads as a contribution to argumenta- tion theory and to the analysis of argumentative discourse. On the other hand, it shows that looking at the roles modal expressions can have in arguments reveals crucial aspects of the meaning and use of the modals themselves. Thus, this book can be read also as a contribution to the semantics and pragmatics of modality. One could swap Modality in Argumentation for Argumentation in Modality as the title of this book. Let us first have a brief look at argumentation and modality, before saying what they have to do with each other and how this book addresses their relationship. 1.1 Argumentation and Modality: Working Definitions In this book argumentation, or argumentative discourse, means a discursive activity in which an arguer puts forward a series of propositions – the arguments or rea- sons – as an attempt to effect the reasonable acceptance of a standpoint by a critic. The standpoint is a proposition to which the arguer is committed and which is not yet accepted, videlicet is doubted, by the critic. The reasonable acceptance is secured by making the standpoint follow inferentially from arguments that are either already accepted by the critic (i.e. are part of the common ground of the arguer and of the critic) or themselves supported as (sub-)standpoints by arguments ultimately grounded in the common ground of the participants. This is not the kind of definition one finds in a logic textbook. It is couched in communicative, dialogical terms and recovers ideas that harken back to the rhetori- cal and dialectical traditions of Antiquity, bringing them together in a way that is 1.1 Argumentation and Modality: Working Definitions 3 typical of a contemporary perspective on argument that is usually characterized as normative pragmatic or pragma-dialectic.4 The domain of modality is vast. It encompasses the concepts expressed by words such as possibility, necessity, probability, but it also covers notions such as ability, need, obligation or risk, and, as we will see, quite a few others. Its linguistic mani- festation is manifold, including both lexical means – ranging from modal verbs to adverbs, adjectives and nouns – and grammatical means, such as grammatical mood and tense. As one could expect with such a broad notional category and so inter- twined with the linguistic systems of human languages, modality has been appre- hended and defined in different ways, with some views considering it a fuzzy category or, even, simply a bundle of related notions chained by family resem- blances and partially sharing their means of linguistic expression (cf. Bybee et al. 1994). Needless to say, entering such a debate would require a book on its own and would lead us astray. Instead, I adopt a concept of modality inherited from a broad philosophical tradi- tion along with a few paradigmatic examples of its linguistic expression. In due course this working notion will be made more precise as I seek to elucidate its rel- evance for understanding how arguments work. Along the way a broader range of – sometimes less obvious – linguistic exponents of the category will be discussed. According to this view, modality is the semantic category associated with the basic human cognitive ability of thinking that things might be otherwise, that is thinking of alternatives: situations other than what is the case.5 Modality refers generally to the linguistic means that allow “one to say things about, or on the basis of, situations which need not be real” (Portner 2009: 1, emphasis is ours). The on the basis of part of Portner’s definition is important because it allows us to recognize the modal nature of examples such as (1): (1) John had to resign. Typically, this kind of English sentence is uttered in a situation where John did, in fact, resign. Most native speakers of English would find a continuation of (1) with but he didn’t unacceptable or at least strange. Yet, even if the event of John’s resig- nation is a fact in the speaker’s world, (1) obliges us to think of the alternatives. It does not mean simply that he resigned, it means that, for some contextually relevant class of alternatives, there was no other possible alternative. There was no viable legal / moral / practical / expedient / stylish, etc. alternative to resignation. Modal 4 The definition offered here is close, in several respect, to the one adopted by one specific norma- tive pragmatic theory, namely Pragma-Dialectics (cf. van Eemeren and Grootendorst 2004: 1). The definition also notably draws on ideas found in Rigotti and Greco-Morasso (2009), Jacobs (2000) and Pinto (1996), especially in what concerns the relationship between argumentation and inference. 5 One prominent logico-philosophical tradition developed the theoretical construct of possible world to deal with reasoning about alternative situations. David Lewis’ (1973) work on counterfac- tuals is one of the most influential in-depth discussions of possible worlds in modern philosophical logic. The book is also noteworthy for the impact it had on the linguistic semantic analysis of modality and conditional constructions. 4 1 Introduction expressions have another important characteristic that (1) showcases: they operate some form of quantification over the relevant alternatives. In (1) we have to do with a universal quantification: all the viable alternatives involve John’s resignation, or equivalently there is no viable alternative that does not involve John’s resignation. It is worth to start noting that as a standpoint (1) requires the support of different arguments than the simple descriptive proposition that John resigned. Being able to show a letter of resignation signed by John would be enough to support the descrip- tive proposition, but not the modal one. Modalities contribute to define argumenta- tively relevant basic semantic classes of propositions, which Freeman (2000, 2005: 93–113), following a rhetorical tradition (cf. Kruger 1975, Fahnestock and Secor 1982), calls types of statement. Rhetoricians have long since observed that there are different broad classes of arguable statements, such as, for instance, evaluative vs. descriptive statements, which require different means of proof when put forth as a standpoint. By contributing to the determine the type of statement, modalities indi- rectly determine what counts as evidence for a given standpoint. This is an impor- tant and often overlooked aspect of the interaction between modalities and arguments. But modalities can be relevant to arguments also in other, more direct, ways, of which something will be said in the following sections of this introduction. 1.2 odality and the Essential Structure of Arguments: M A First Reconnaissance How relevant is modality for argumentation theory? Obviously, it will be the task of the entire book to provide answer to this question, but we need a provisional one to justify that the journey is worth taking. Consider the following dialogue: (2) A: I heard that Suzy got caught shoplifting yesterday afternoon. B: She can’t have done it. She and I were at the movies all afternoon. The turn of speaker B contains an argument: the standpoint that Suzy is innocent of shoplifting yesterday afternoon is supported by the alibi that Suzy was at the movies with speaker B all afternoon. The turn of speaker B also contains a modal (can’t), which refers to some kind of impossibility. Intuitively we recognize that this notion of impossibility plays a role in the argumentation. But what role is it? The modal appears in the sentence that manifests the standpoint defended by B, but it is not entirely clear whether the modal should be considered part of the content of that standpoint: should we consider that B was defending the standpoint ‘It is impossible that Suzy got caught shoplifting yesterday afternoon’ or the standpoint ‘Suzy wasn’t caught shoplifting yesterday afternoon’? If the modal is not part of the content of the standpoint we might consider the hypothesis that it signals something like the force of conviction with which B upholds the standpoint, or the force that the 1.2 Modality and the Essential Structure of Arguments: A First Reconnaissance 5 1. It is impossible (=can’t) that Suzy got 1. Suzy wasn’t caught shoplifting caught yesterday afternoon shoplifting yesterday afternoon [standpoint] [standpoint] can’t [modal] 1.1 She and I were at the 1.1 She and I were at the movies all afternoon movies all afternoon [argument] [argument] Fig. 1.1 Structural graph representations of the alternative analyses of example (2) argument lends to the standpoint. The alternative analyses can be visualized as in Fig. 1.1, employing a style of representation commonly used in argumentation the- ory (cf. Snoeck Henkemans 1997; Freeman 1991, 2011; Vorobej 2006).6 For instance, we might compare the argument in (2) and the one in (3), below, and observe that the former is presented as making a stronger case for Suzy’s innocence. (3) A: I heard that Suzy got caught shoplifting yesterday afternoon. B: That’s unlikely. She’s not that kind of girl. We might as well consider that the impossibility manifested by can’t has to do not only with the force but also with the kind of argument used, with the precise way in which the argument presented works in supporting a conclusion: the alibi works as an argument because it’s physically impossible for people to be in two places at the same time. Then the analysis would look quite different from either option in Fig. 1.1. Figure 1.2, provides a rough sketch of what an analysis of the argument along these lines would look like. In this case we would say that the modality expressed by can’t has to do with what Ancient and Medieval theorists called the topos or locus from which the argu- ment is drawn (Rigotti and Greco-Morasso 2010). Some modern argumentation theorists here would prefer to speak of argument schemes (van Eemeren 2010) or argumentation schemes (Walton et al. 2008). Now consider the substitution with another modal expression, certainly, as shown in (4). (4) A: I heard that Suzy got caught shoplifting yesterday afternoon. B: Certainly she did not do it. She and I were at the movies all afternoon. 6 The specific type of graph representations used in the book derives from the view of argumenta- tion structure espoused by the Pragma-Dialectical theory of argumentation (cf. Snoeck Henkemans 1997; van Eemeren et al. 2007, Chapter 7). 6 1 Introduction 1. Suzy wasn’t caught shoplifting yesterday afternoon [standpoint] 1.1 It is impossible (=can’t) that Suzy got caught shoplifting yesterday afternoon 1.1.1.a (It’s physically 1.1.1.b She and I were at the impossible for people to be in movies all afternoon. two places at the same time.) [argument] Fig. 1.2 A different graph structure for the reasoning in example (2) Leaving aside, for the moment, finer semantic and pragmatic considerations that can explain why certainly feels somewhat awkward in B’s answer, let us focus on one feature of the modal certainly: it seems to tell us about how strong a case the arguer is making (a rather strong case), but, contrary to what happens with can’t, we do not feel that the modality is also pointing to the kind of argument used. Words like certainly are often said to express a doxastic modality, a modality concerned with the alternatives that are admissible within the belief system of a given agent or, more simply, the alternatives consistent with the beliefs of that agent. Another term frequently used is epistemic. The term epistemic is however ambigu- ous between a broad reading where it is coextensive with doxastic and a strict read- ing where it refers only to what is actually known, rather than simply believed. The adverb certainly directly refers to the arguer’s mental state of ‘certainty’ and con- veys the intended strength of the argument via the reference to the arguer’s subjec- tive strength of conviction. What certainly does not show is that the arguer’s belief in Suzy’s innocence derives from a special kind of impossibility. Consider now another variant of our example, presented in (5), where the semantic contribution of the modal also diverges from (2), but in a direction that seems quite opposite to what we have in (4). (5) A: I heard that Suzy got caught shoplifting yesterday afternoon. B: She could not do it. She and I were at the movies all afternoon. The difference between could not do and can’t have done is easily perceivable but difficult to pinpoint. In order properly draw this distinction one needs theoretical machinery that will be introduced in the following chapters (see, in particular, Chap. 5, § 5.2.4.4). But there are certain informal remarks that can enrich this preliminary illustration of how the meaning of modal expression can be finely intertwined with the structure of arguments. 1.2 Modality and the Essential Structure of Arguments: A First Reconnaissance 7 We can start by observing that in (5) the temporal operator takes scope over the modal, while in (1) it is the other way round. In (5) the impossibility of shoplifting is placed in time: there was a time when it was impossible for her to shoplift. In (2), on the contrary, the impossibility that she shoplifted in the past is seen as present and, perhaps, eternal. The speaker abstractly considers that the proposition ‘Suzy shoplifted yesterday afternoon’ is incompatible with another proposition which is a known fact (‘She and I were at the movies all afternoon’), given a certain accepted ontology of the physical world. The temporal version in (5) brings in an idea of causality, which is absent from (2): being at the movies is not just something that is incompatible with shoplifting but something that prevents it from happening. In real discussion contexts this might trigger additional conversational implicatures. It might suggest that Suzy is somewhat inclined to that sort of vice, and she could have given in to that kleptomaniac inclination, weren’t it for the particular situation that prevented her from doing so. The doxastic modality in (4) is about the relationship between the proposition and speaker’s internal belief states, while the modality in (2) concerns the logical incompatibility between the proposition and known facts in a certain ontology of the physical world – some scholars would characterize it as simultaneously alethic and epistemic. It concerns the relationship of the proposition with certain facts of the world, namely the world’s physical make-up. It is therefore an ontological modality. Another, more widespread, term for these modalities is alethic,7 as facts are propositions that are the case in the world. Since these facts are known facts, the modality can be characterized also as epistemic, both in the broad and in the strict sense. Example (5) is about the relationship between the proposi- tion – or perhaps the event it denotes – and the circumstances – some students of modality would call it circumstantial, but, again, some would also insist it is alethic/ ontological too, as circumstances are facts.8 If we look at it from the viewpoint of argumentation, this last kind of modality seems to be primarily about the real-world relation that underlies the argument (a circumstance preventing an event), second- arily about the corresponding logical relations (the incompatibility between the two propositions), and only in a very indirect way about the beliefs of the speaker (his/ her dismissal of the proposition). This book adopts the hypothesis that all these levels of consideration plus a further one – speech acts – are relevant for understand- 7 The term alethic modality (from Gr. alétheia ‘truth’) is widely used in the philosophical and logi- cal literature (but less so in linguistics) to refer to what is possible or necessary in view of true facts. Different kinds of alethic modality can be envisaged according to the range of facts that are taken into consideration: physical, metaphysical, logical, etc. Alethic modalities will be examined in detail in Chap. 5 as part of a semantic typology of modalities geared at capturing their role in arguments. 8 On the vast range of everyday alethic modalities see the brilliant remarks in Lycan (1994: 171– 178). Note that among certain students of linguistic semantics (cf., for instance, Lyons 1977: 791; Palmer 1986: 11; Papafragou 2000; Nuyts 2006: 8) alethic modality is understood in a much nar- rower sense as synonymous with logical modality. For a more thorough discussion of varieties of alethic modality among the linguistically and argumentatively relevant “flavors” of modality the reader is directed to Chap. 5. 8 1 Introduction ing how arguments work and modal expressions, of the different kinds, have an interesting story to tell about them. For the moment it is not necessary to push the analysis of these examples further. The main point it was meant to make is simple: the argument in (2) does not simply happen to refer to impossibilities just like it happens to refer to shoplifting and going to the movies. Rather, can’t and the other modals examined appear closely con- nected to the essential properties of the argument. That is to what makes it an argu- ment and to what makes it an argument of a certain kind. This book thus examines modality in relation to the essential structure of arguments. 1.3 Aims of the Study The exploration of modality’s relation to the essential structure of arguments will be guided by two related questions. These questions were first raised by Stephen Toulmin and have thence lingered under various guises in argumentation theory without finding a satisfying answer. (Q1) How are modals similar to arguments? In the first essay of the Uses of Argument not only Toulmin argues that modals are best understood by looking at their role in arguments, he goes on to establish if not a complete identity, at least an isomorphism between the structure of modal mean- ings and the structure of arguments, distinguishing in both modal meanings and argumentative moves an invariant force and a variable, contextual, set of criteria depending on the field to which the propositions concerned belong. I take up from Toulmin the question of the relationship between the semantic analysis of modality and the structure of arguments. At first sight, it may seem a most abstruse question, at least from the viewpoint of someone who is mostly interested in the analysis and criticism of actual arguments, rather than in fine points of semantics. In fact, it turns out to be a very productive question to look into. (Q2) Is there a special place for modality in the structure of arguments? In the third essay of the Uses of Argument – the most widely cited in the argumenta- tion literature – Toulmin finds a special slot for modalities, the qualifier in his analy- sis of the fundamental “layout” of arguments – the eponymous Toulmin model. This slot is distinct from both the standpoint and the argument.9 The graph representa- tions of arguments used by argumentation theorists, which we introduced in the previous section, derive, in part, from Toulmin’s own view of argument layout. Toulmin sees the qualifier as an indication of the strength of the inferential step 9 In the eponymous Toulmin model the (modal) qualifier, indicating the strength of the inferential step, is a slot of its own and is neither part of the claim (that is the standpoint) nor of the data that support it. 1.3 Aims of the Study 9 from the argument to the conclusion.10 The special place of the qualifier in the Toulmin model brings us to the question of the distinctive place, or role, of modality in the fundamental structure of arguments. This is perhaps the question that argumentation scholars have more explicitly considered. Students of argumentation have typically dealt with the role of modality in the analysis of arguments in two ways. Either modality is seen as an indication of the strength of support that the premises offer to the conclusion (cf. Freeman 1991, 2011) or alternatively, in pragmatic terms, as an indication of the degree of commit- ment of the arguer towards the propositional content of the standpoint being advanced (cf. Snoeck Henkemans 1992; Tseronis 2009). In this book this second question will be examined and reframed in the light of the answer that will be given to the first question. To these two Toulminian questions I add a third one, which is best answered not by considering modality generally, as a semantic category, but by looking closely into the linguistic semantics of specific modal expressions in a given language. (Q3) What does the use of modal expressions in an ordinary discourse tell us about the arguments being put forth in the discourse? This question bridges research on argumentative indicators (van Eemeren et al. 2007) in argumentation theory with current research on the discourse semantics of linguistic constructions. An argumentative indicator is a word or expression that can tell us something important about the argumentation manifested by a text or dia- logue. Argumentative connectives marking an utterance either as an argument (e.g. because, since) or as a conclusion (e.g. therefore or so) are perhaps the simplest example of argumentative indicators. Yet, the notion of indicator is much broader than that, and the indication could concern any kind of information relevant for reconstructing how an argument works. The notion of indicator is intimately tied to the idea that the superficial verbal form of arguments is not immediately transparent and that their fundamen- tal organization needs to be reconstructed in order to be able to evaluate them. Research on indicators takes the perspective of the analyst for whom reconstruc- tion is a necessary step of argument criticism. As a consequence, the research on indicators takes a rather practical stance considering linguistic structures: whatever can help the analyst to establish with reasonable probability the commitments of the arguers can be an indicator. As observed in Rocci (2008: 166), the notion of indicator is not restricted to expressions that refer to arguments. Often an indicator does not refer to the argumentative move itself but carries some other semantic, pragmatic or even stylistic feature, which, for a variety of reasons, can happen to be often— not necessarily always —concomitant with that argumentative move. So, a verbal expression can be an indicator per accidens because of its significant correlation with the performance of an argumentative move. I call these indicators 10 More precisely the qualifier indicates the strength conferred by the warrant to the inferential step from the data to the claim (Toulmin 2003: 94). The place of modality in the Toulmin model will be discussed at length in Chap. 3. 10 1 Introduction indirect (cf. Rocci 2008: 167). As long as we deal with indirect indicators the point of view of the analyst and that of the communicators may partially diverge. On the one hand, the indicators will be, in every case, based on the meanings to which the arguers are committed. In interpreting these meanings the analyst will rely on “linguistic knowledge” and on “pragmatic knowledge of how what is said should be interpreted in the context” (Snoeck Henkemans 1992: 103). This is the linguistic and pragmatic knowledge that is involved in the normal process of mean- ing construction and interpretation, as stressed by Snoeck Henkemans (1992: 103): Arguers who are engaged in a critical discussion aimed at resolving a dispute must observe the pragma-linguistic conventions for appropriate use of the words and expressions they use in defending theirs standpoints. Since the analyst may assume that the arguer has observed these conventions, he will attempt to give an interpretation of the arguer’s standpoint11 that is plausible in the light of the conventions. Yet, in the case of indirect indicators, once the meanings of the arguers have been thus clarified, the analysts use these meanings as indirect evidence for recon- structing the argument, following practical rules that are not necessarily part of the normal process of discourse understanding. For instance, van Eemeren et al. (2007: 172–173) observe that the use of future tense markers (will, shall, be going to, etc.) is an indicator that some type of causal argumentation scheme is being used to support the standpoint. In reality, the link between the use of the future and the argument scheme in question is indirect and connected with different forms of causal reasoning in different ways. On the one hand, statements in the future tense are often predictions, and predictions are often justified causally by arguments from cause to effect—such as in (6.a) below. On the other hand, deliberation on future actions is often justified by the positive/ negative nature of the (later) future consequences of these actions. In this second case both the standpoint and the argument refer to the future, and the inferential link goes from the evaluation of the effect to the evaluation of the cause—as shown in example (6.b). (6.a) Arg: Storms have devastated the crops throughout the country. Stp: The price of vegetables will increase. (6.b) Stp: Travelling with Mark is definitely a bad idea. Arg: He will annoy you to no end with his strange fixations about hygiene on buses and airplanes. The correlation between the use of the future and causal reasoning is a “rule of thumb” that analysts can profitably follow, but it does not correspond to one mean- ing construction and meaning interpretation path by the arguers. There are, however, words and grammatical constructions whose business con- sists, at least in part, in telling something about the arguments being put forth in discourse. I call these linguistic structures direct argumentative indicators. In this book I investigate the hypothesis that at least some modal constructions function semantically as fine grained direct indicators of argumentation, guiding the In the context of Snoeck Henkeman’s original discussion only the reconstruction of standpoints 11 was at issue, but the same could be said of the reconstruction of arguments. 1.3 Aims of the Study 11 addressee towards the intended interpretation of argumentative connections between utterances in discourse. It is clear, though, that by saying that modal expressions tell something about arguments, I do not mean that they are like meta-argumentative expressions, such as premise, conclusion, valid, deductive, concession, and so on. One could say they are epi-argumentative, somewhat akin to performative expres- sions and even more to illocutionary modifiers, in that they flag the very utterance in which they appear as having a given argumentative function or guide the infer- ences of the interpreter by restricting the range of possible interpretations. One key difference between these direct indicators and the most classic examples of performative and illocutionary modifying expressions is that they do not concern the pragmatic function of the utterance taken in isolation – if ever there is such a function – but contribute to clarifying how the utterance contributes to a coherent discourse conveying an argument. A second difference concerns the way in which the conventional linguistic meaning of the direct indicators interacts deeply with finely specified types of contextual information that the addressee is invited to recover from the ongoing discourse or situation. Contrary to what happens with the indirect indicators, here the perspective of the analyst, the perspective of the addressee and the perspective of the communicator trying to constrain the interpretation process of the addressee are tightly coupled. Understanding the indicator is part and parcel of normal discourse understanding rather than a specialized activity of the analyst. This makes direct argumentative indicators a particularly precious resource in view of a pragmatically realistic recon- struction. In this book I will typically adopt the perspective of the addressee, assum- ing that what is clearly relevant for the natural understanding of an argument as an argument is, a fortiori, relevant for its reconstruction and subsequent criticism. Direct indicators are also the terrain where the research on argumentative indica- tors can meet a large body of recent research on the semantics of discourse under- standing centered on the contribution of the conventionally encoded meanings of linguistic constructions to the understanding of broad stretches of discourse. Bringing together these two perspectives is an intended methodological contribu- tion of this book. Not surprisingly, in the argumentation literature the role of modals as indicators has been mostly tied to the speaker’s indication either of the strength of support that a standpoint enjoys or of the degree of commitment towards its content. Some authors have gone further taking the indication of the strength of support as an indi- rect clue for understanding argument (macro-)structure, that is how the different reasons that make up an extended argumentation fit together to lend support to the conclusion (Cf. Freeman 1985, 2011; Snoeck Henkemans 1992). As for the modal expressions taken into consideration, the modal adverb probably is the only one that received several fully fledged semantic analyses by argumentation scholars (cf. Toulmin 2003; Ennis 2003; Pinto 2007; Tseronis 2009). Through the semantic analysis of different modal expressions in English and in Italian this book will show that this focus on force and degrees of commitment pres- ents a far too restrictive view of modals as argumentative indicators, and that there is definitely more that they can tell us about the workings of arguments. Again, the 12 1 Introduction strategy for the analysis of modality developed in answering our first question will serve us well in exploiting modals as indicators. The present investigation on modal constructions as direct argumentative indica- tors will both benefit from and contribute to a relatively recent surge of research on the semantics of modality in relation to contextual and discursive information. The book draws extensively on the theories and techniques developed in linguistic semantics – in particular in the tradition of formal semantics – for the analysis of modal meanings as well as on the author’s own work on the analysis of modal meanings in discourse context. In addressing the three questions outlined above, I will keep a broader purpose in mind. I want to provide an illustration of the proper place of semantic consider- ations in argumentative analyses, and particularly in those analyses which seek to capture the dynamics of argumentative discussions in the concrete contexts of social life in which they arise. As it discusses matters that are of interest for semanticists and discourse analysts, the book remains anchored to the perspective and concerns typical of argumentation theory: the analysis of ordinary arguments oriented towards their critical evaluation. The task of Chap. 2 will be precisely to focus on the specific disciplinary viewpoint of argumentation theory on the analysis and evaluation of arguments and make explicit the role that semantic and pragmatic considerations can play from that specific viewpoint. 1.4 A Twofold Research Strategy An analogy could help clarifying further the perspective adopted by this book: modality is investigated here in relation to argumentation very much like the notional semantic category of time could be investigated in relation to narrative. That is by asking how temporal relations between states and events – as expressed by tenses, temporal adverbials and other means – contribute to make a narrative discourse what it is.12 With modality and arguments things are however more com- plex than the analogy would suggest. The discussion of our initial set of examples already showed that modality comes in different “flavors” – as semanticist von Fintel (2006) would put it. We have already distinguished alethic, epistemic and doxastic modalities and observed that there are different sorts of alethic modalities. This book will develop the hypothesis that these different kinds or “flavors” make a distinct contribution with what I have called the essential structure of arguments. 12 The comparison with the role of time in narrative can be more telling if we consider that narrativ- ity requires temporality – a text that does not order events in time cannot be a story – but temporal- ity does not represent a sufficient condition: it is not enough to order events in time to tell a story. For instance, narrative relations have a pragmatic facet, they are an action of a narrator telling a story to the narratee, and the picture can be further complicated by the presence of a point of view in the narrative and by the evocation of multiple voices, which aren’t identified with the narrator nor with the narratee. 1.4 A Twofold Research Strategy 13 Providing a clear definition of these “flavors” and, even more crucially, explain- ing how they are obtained in the interpretation of an utterance is a crucial task for a linguistic semantics of modality. In order to account for the fact that many modal expressions, such as modal verbs or adjectives, can manifest according to their dis- course context a seemingly boundless variety of modal “flavors”, the formal seman- tics tradition has developed a treatment of modal semantics as relational and context dependent. This theory, which goes under the name of Relative Modality was first proposed by the German linguist Angelika Kratzer (1977, 1981, 1991), in a close dialogue with the contemporary logical and philosophical work on possible worlds semantics that was being developed by David Lewis ( 2007). This theory forms the starting point of virtually all the work currently being done on modality in formal semantics (See von Fintel 2006; Kaufmann et al. 2006 and, in particular, Portner 2009 for recent syntheses of this body of work) and has inspired more infor- mal linguistic work on modality, especially in relation to the “interface” between linguistic meanings and pragmatics (cf. Papafragou 2000). The central idea of this theory is that the modals are relational predicates of the form M (B, p). They take (at least) two arguments: the propositional content in their scope – the prejacent13 proposition p – and a set of propositions, called the conver- sational background (B). Modal expressions of necessity like the modal auxiliary must can be understood in terms of the logical consequence of the prejacent from the conversational background, while possibility expressions such as the modal auxiliaries can or may are to be conceived in terms of the logical compatibility between the prejacent and the background: ‘necessary’/ ‘must’ (B, p): p is a logical consequence of B ‘possible’/ ‘can’/ ‘may’(B, p) : p is logically compatible with B Assuming a possible world reading of logical consequence, saying that p is a logical consequence of B is equivalent to saying that p holds true in all the alternative worlds or situations where the propositions of the set B are true. While, in the case of logical compatibility, p will be true in at least one of the worlds or situations where the propositions of B are true. For instance, can’t in example (2) discussed above could receive an interpretation along the following lines: ‘Suzy was shoplifting yesterday afternoon’ is incompatible with (what is known of the whereabouts of Suzy yesterday afternoon & what is known about the functioning of the physical world) One advantage of this kind of approach – the one that immediately caught the atten- tion of linguists – is that the variety of modal “flavors” results from an invariant modal force and a variety of conversational backgrounds reconstructed in the con- text of utterance. There are however other interesting advantages of this general approach that began to be noticed. One of them is that it offers a way of describing how modal meanings relate to other information in a developing discourse (cf. Roberts 1989; Geurts 1999). 13 I follow von Fintel (2006) in recovering this term introduced by medieval logicians. 14 1 Introduction This book will develop the idea that this style of semantic treatment of modality is not only successful in its original endeavor of accounting for the different modal flavors but provides us with the basic tools for explaining the interaction of modality with discourse context. In particular, the relational and contextual approach will allow to investigate how the relational semantic structure of modal meanings con- tributes to define argumentatively relevant relations in discourse at different levels including “real world” relations, logical relations, inferential relations and speech- act level relations. The investigation presented in the book also adopts a second perspective on the relationship between modality and argumentation, which complements the one out- lined above. To use again the time-narrative analogy: once it has been established that in a language – say, French – grammatical tenses are one of the main means for expressing time relations, one could examine more closely the use of tenses in nar- ratives, perhaps to discover that their business there is not limited to expressing time relations and that they contribute also other information – perhaps somewhat related to time, perhaps arbitrarily packaged together with it – which is important for mak- ing a story what it is. In other words, with reference to a specific language, the meaning of the grammatical morphemes of tense can be seen, in part, as contribut- ing directly to the discursive construction of narrative, without the mediation of a referential category such as time. An extreme, not necessarily advisable, example is represented by Harald Weinrich’s (2004 ) well known attempt to characterize the contribution of French tenses to narrative and to other modes of discourse completely dispensing of time reference.14 Similarly, in connection with the third research question of the book constructions based on modal verbs and mood mor- phemes will be examined here for their full potential as direct argumentative indica- tors. Any way in which they shape or constrain the interpretation of the arguments in which they appear is relevant here, without limiting the observations to those aspects of an argument – such as strength of support – that seem more intuitively or more directly connected with the concept of modality. In doing so, we can end up learning more about modality and related notions and refine our methodological tools for argument criticism. Drawing from the terminology of linguistic semantics, we may call the perspec- tive moving from the notional category of modality onomasiological and the second perspective, which investigates the argumentative functions of linguistic forms, semasiological. One of the main findings of this second research strategy in the book is that epis- temic and doxastic modal expressions are found to be particularly revealing of the 14 In the strongly anti-referentialist structuralist climate of the early 1960s, Harald Weinrich (2004 ), a student of French linguistics, famously set out to create a structuralist text linguistics that sought to connect the structure of narrative directly with tenses, dispensing of the troublingly referential notion of time. In the process Weinrich discovered interesting non temporal properties of French (and Italian) tenses, which he (ironically) conceptualized in terms of different attitudes of the emitter (narrating vs commenting) towards the (reference!) world. Later on, Weinrich had to soften considerably his once belligerent stance on the irrelevance of time, while keeping the other interesting aspects of his analysis. 1.4 A Twofold Research Strategy 15 functioning of the arguments in which they appear because they typically also con- vey specific evidential meanings. Evidentiality is the semantic category correspond- ing to the indication of the speaker’s source of knowledge (evidence) of the propositional content of the utterance (cf. Chafe and Nichols 1986; Dendale 1994; Squartini 2004, 2008). In titling an early and influential collection of studies Chafe and Nichols (1986) evocatively characterized evidentiality as “the linguistic coding of epistemology”. In sharp contrast with the bi-millennial scholarly history of modality, the notion of evidentiality originated fairly recently from a field of inquiry far removed from the logical and philosophical analysis of language. The idea first emerged in the grammatical description of certain native American languages where the grammati- cal marking of the information source – typically distinguishing direct perception, hearsay and inference – is obligatory to produce a grammatically acceptable sen- tence. The notion and term were then used in typological linguistics to refer to the (many) languages of the world where such a grammatical system exists (Guentcheva 1996; Aikhenvald 2004). However, English and the other Western-European lan- guages are not among those where evidentiality is fully grammaticalized. Possibly because of its disciplinary origin and of the scarcity of studies on the evidential strategies of English and of other major European languages, the analysis evidentiality was absent for a long time from theoretical discussions in semantics and in the philosophy of language. Its impact on argumentation studies was non- existent, despite the obvious connection between the linguistic marking of types of evidence in the utterance and the discursive presentation of the evidence itself in the form of arguments. It is only recently that indicators of the source of evidence have become the object of focused attention in semantics, also in European languages (such as English, Spanish, French or Italian) were their presence is not obligatory (cf. Dendale and Tasmowski 1994; Aijmer 2009; Cornillie 2007; Pietrandrea 2007; Squartini 2004, 2008) and their close connection with the semantics and pragmatics of epistemic and doxastic modality has begun to be explored in theoretical studies (cf. Faller 2002; McCready and Ogata 2007; Sbisà 2014). Remaining focused on the central notions of the category of modality such as necessity and possibility, this book offers an in-depth discussion of the conceptual relationship between evidentiality and argumentation and a detailed analysis of how the evidential component in the semantics of several markers of modality interacts with argumentation.15 15 In fact, the line of research developed in the present book was the inspiration of a closely related but autonomous research focusing specifically on evidential constructions as direct argumentative indicators carried out by Johanna Miecznikowski and Elena Musi within the research project From perception to inference. Evidential, argumentative and textual aspects of perception predicates in Italian (Swiss National Science Foundation Grant n. 141350). See, in particular, Miecznikowski and Musi (2015), Musi (2014, 2015). 16 1 Introduction 1.5 Focus on Italian Linguistic Structures A and on Contextualized Discourse Data It is clear from the above remarks that this kind of fine grained investigation of direct argumentative indicators needs to be carried out with reference to the seman- tics of given linguistic units in a target language, a language well-known to the analyst, ideally his/her native tongue. It also needs to be carried out using authentic data considered in their rich context, and this context needs to be reasonably known to the analyst so that it can be unraveled when reconstructing underlying reasoning requires so. In this book the languages investigated are Italian, and, to a lesser extent, English. While the theoretical discussion of modality in relation to the essential structure of arguments will mostly exploit English examples, both invented and selected ad hoc from various sources, the study of modal markers will be largely devoted to the Italian modal system, bringing together the results of several previous semantic studies, leveraging on my native intuitions and exploiting systematically rich corpus evidence. So, as we move from English to Italian data through the book, we also move from out of context illustrative examples to examples extracted from a large care- fully designed corpus, representative of a family of closely related discourse genres and of a given context of social interaction. The corpus, which represents the main source of Italian language data for this book, was collected for the research project Modality in Argumentation. A semantic-argumentative study of predictions in Italian economic-financial newspapers.16 The corpus amounts to roughly 4 million token words and consists of journalistic articles collected from three specialized Italian economic-financial dailies (Il Sole 24 Ore, Italia Oggi and MF/Milano Finanza). On the one hand, the extensive discussion of Italian expressions of modality and the analysis of samples of argumentative discourse in Italian are important features of this work, and I sincerely hope they will be of interest also for linguists working on the semantics and pragmatics of Italian modal expressions. On the other hand, this decision may seem to restrict the appeal of the final part of the book to scholars working with Italian or other Romance languages. In fact, I believe it is not necessarily so: the book is accessible to a broad range of readers and remains, first of all, a semantic contribution to the core of argumentation theory. Readers unfamiliar with Italian will be perfectly able to follow the analyses through the English translation of the examples and, where necessary, the glosses. They will be lead to compare and contrast them with parallel English expressions. More explicit comparisons with English and other languages will be drawn when relevant, with pointers to the literature. 16 The project Modality in Argumentation. A semantic-argumentative study of predictions in Italian economic-financial newspapers was supported by the Swiss National Science Foundation (Grant: 100012-120740/1) from September 2008 to December 2011. See also footnote 1. 1.6 What This Book Is Not About 17 This will not take completely away a feeling of strangeness that readers unfamil- iar with Italian will get from the final parts of the book and the impression that some of the underlying theoretical distinctions may well be language specific and that they may turn out not to be extensible or useful in approaching argumentative dis- course in English or in other languages. This is, however, a healthy feeling. This is an often ignored caveat that all argumentation scholars should bear in mind when they go about developing conceptual distinctions on the sole basis of analyses of English argumentative discourse. At the same time, if the fine analyses of the English adverb probably which have been offered by Toulmin and by other authors reviewed here have been rightly considered not just as accounts of peculiarities of a curious English idiom,17 but as insightful discussions of the role of modality in argu- ment, there is no reason of principle for considering accounts of Italian modal verbs or adverbs more parochial. If anything the parallel consideration of English and Italian found in several passages of this book (with the occasional nod to French) should sharpen our ability to capture general underlying dynamics, even if an explicit and systematic comparative investigation of modality in the two languages lies well beyond the scope of the book. The journalistic corpus employed as the main source of data allows to focus on the role of modality in arguments within a given context of social interaction, namely the media coverage of the economy and, more specifically, of the financial markets. The study of argumentation in context, a burgeoning area of argumentation stud- ies, appears prima facie far removed from issues of semantics. In fact, it is not so. In Chap. 2 I will argue that semantic considerations, in general, can play an impor- tant role in understanding how arguments relate to their social contexts, and modal- ity, in particular, is systematically tied to the dynamics of contextualized activity types, to the point that it can function as a highly sensitive contextualization cue (cf. Gaik 1992) for argumentation. More specifically, Chap. 6 will provide a short over- view of the broader dynamics of financial communication to which the business- financial journalistic coverage participates (cf. Rocci 2014; Rocci and Luciani 2016; Westbrook 2014) highlighting the correlation between the modal semantic struc- tures deployed in that discourse genre and a number of social, physical and epis- temic constraints of the underlying field of social interaction. 1.6 What This Book Is Not About Modality and argumentation are two hugely vast domains: investigations into each of them span over two millennia, across a range of disciplines encompassing phi- losophy, logic, rhetoric and linguistics, with modern ramification into artificial 17 It is worth noting how, at the time of publication of The Uses of Argument, the allegation of play- ing otiosely with English idioms as hobbyist lexicographers was a common (and for the most part unfair) accusation leveled against ordinary language philosophers by more logically oriented ana- lytic philosophers. 18 1 Introduction intelligence and the mathematical theory of probability. The aims and limitations of the present study will emerge more distinctly once we have cleared the way of a few big issues that lay at the intersection of the two vast domains, which are not addressed directly in the book. These issues are not attacked up-front, yet they will continue to lurk in the background to surface again at various points in the thread of the main argument of the present work. Thus, I take the liberty of evoking briefly some of the issues the present book is not directly about. 1.6.1 This Book Is Not About Modal Logic It should be said from the outset that this book is not primarily about the logic of modalities. The specificities of arguments containing modal propositions have been addressed since Aristotle and modal logic, understood as the normative study of the validity of arguments containing modal propositions, flourished particularly in the Middle Ages, to be later almost forgotten and emerge again in the Twentieth Century in the context of formal logic. For the purposes of the present research, modal propositions can be defined as complex propositions formed by a modal operator taking a prejacent proposition in its scope. This is a modern view of modality. Antiquity and, more explicitly, the Middle Ages viewed modality under many respects equivalently, but through a dif- ferent set of categories and defined modal propositions accordingly. The anony- mous treatise De propositionibus modalibus (cf. Anonymous 2006), traditionally attributed to St. Thomas Aquinas, defines a modal proposition as a proposition where a “mode” (i.e. a “determining attribute”) is attributed neither to the subject of the proposition (as in Socrates is white) nor to the predicate (as in Socrates runs well), but to “the composition itself of the predicate with the subject, as when it is said ‘that Socrates runs is impossible’” (cf. Uckelman 2009b: 2 for the English translation). In chapters 12–13 of the book On Interpretation (cf. Hintikka 1973: 41–61) Aristotle captures the immediate inferences (conversion and opposition) that can be established with regards to propositions constructed with modalities and negation.18 As nicely put by Hacquard (2011: 1486), “under all types of interpretations, possi- bility and necessity modals enter into patterns of entailments and logical equiva- lences similar to those involving universal and existential quantifiers”. As it is well known, since late Antiquity, these patterns have been beautifully mapped onto the traditional square of opposition (cf. Horn 2001 , 2014) (Fig. 1.3). Contradictory propositions, connected by the diagonals of the square, entail each other’s falsity, while they are, in turn, entailed by the negation of their opposite. According to the respective scope of the negation and the modal, the negation can be either inter- 18 nal (when the negated proposition forms the prejacent of the modal) or external (when the whole modal proposition is negated). 1.6 What This Book Is Not About 19 Fig. 1.3 The modal square of opposition Thus, contradictories are logically equivalent (=) to the negation of their opposite on the diagonal of the square: (7) Possible p = Not impossible p (8) Necessary p = Not possible not p As noted by Hacquard (2011) cited above, these inferences remain valid across all flavors of terms such as possible and necessary and across a wide range of isomorph modal concepts. This includes not only the alethic and epistemic modalities that have been introduced in the previous sections, but also deontic concepts, corre- sponding to those moral, legal or practical alternatives that were evoked in com- menting the necessity of John’s resignation in example (1). Let us take obligatoriness (a deontic necessity) and permissibility (a deontic possibility): from the obligatori- ness of wearing a bike helmet I can deduce that it is not permissible not to wear one, and vice versa. From the permissibility of smoking in the dedicated smokers’ areas of the airport, I can deduce that smoking is not forbidden there. Contrary propositions on the upper corners of the square do entail each other’s falsity. Yet, there are not equivalent to the negation of their opposite. Traditionally put, contraries cannot be true at the same time, but can be false at the same time. (9) Necessary p ⇒ Not impossible p (10) Impossible p ⇒ Not necessary p To exemplify: from the obligatoriness of helmets I can infer that they are not forbid- den (9), but I cannot infer obligatoriness from the lack of a prohibition. From the prohibition of smoking I can infer that it is false that smoking is mandatory (10), yet from knowing that it is false that smoking is mandatory, I cannot infer that it is 20 1 Introduction indeed prohibited. Additionally, the propositions in the upper corner entail their subaltern propositions in the corresponding lower corner: (11) Necessary p ⇒ Possible p (12) Impossible p ⇒ Possible not p If I know that helmets are compulsory, I can immediately infer that they are also allowed (11), and if I know that smoking is forbidden on the restaurant’s premises I can infer that it is not mandatory (12). All these inferences are deductively valid and are immediate, requiring only one premise rather than the composition of two or more statements. As it is well known, the immediate inferences enshrined by the square assume that modal expressions such as Possible p can refer also to states of affairs that are necessarily the case and not only to those that are merely contingent. In fact, Aristotle saw possibility expressions such as Greek dunatón ‘possible’ and endechómenon ‘contingent’ as “homonymous” (cf. Prior Analytics I, 3, 25a37–40), i.e. as being ambiguous between two readings, so that he recognized two distinct notions of possibility, albeit without providing a distinct term for each (Hintikka 1973: 29). We can speak of possibility proper or unilateral possibility for the notion captured by the square, and of bilateral possibility (see also Horn 2001 ) for the notion that is incompatible with necessity. The term contingent (Lat. contingens) started to be specialized to indicate bilateral possibility in the works of Late Medieval logicians (cf. Knuuttila 1982). Modern linguists have offered pragmatic explanations of the two readings of possibility expressions (Horn 2001 ; van der Auwera 1996), according to which the unilateral meaning is basic and the bilateral meaning arises as the result of a Gricean pragmatic inference (a scalar Quantity implicature). The utterance of the weaker, less informative, proposition Possible p invites the hearer to conclude that the stronger, more informative, proposition Necessary p is not the case, unless there is evidence of the contrary. Thus, a layer of defeasible pragmatic inferences is superposed to the logical immediate inferences captured by the square. Neither the immediate inferences captured by the square, nor the scalar pragmatic inferences superposed to them occupy the central stage in the present work. In contrast, much will be said about the systematic ambiguity or, better, under- specification of many modal expressions between the different flavors (e.g. alethic, epistemic, deontic), about the pragmatic and discursive processes invoked to resolve it. In this context, a little will be also said about the logical and pragmatic inferences that “bridge” different flavors. We have already seen, in Sect. 1.2, that alethic impos- sibility pragmatically implicates epistemic impossibility. It also pragmatically implicates doxastic impossibility.19 For instance, the deontic obligation to perform an action presupposes that the action is alethically possible in a very specific sense, i.e. that it is doable by the intended subject of the obligation. On the other hand, in 19 This is a quality implicature, as one cannot coherently say that a state of affairs is alethically impossible and then admit it as a belief. Doing so would amount to presenting oneself as either insincere (I say what I don’t believe) or irrational (I admit beliefs that I know to be impossible). 1.6 What This Book Is Not About 21 many contexts the assertion of an alethic possibility pragmatically implicates also an epistemic uncertainty (two-sided possibility), via a bridging quantity implica- ture, as in (13). (13) They can win the match The assertion of (13), in most contexts implies that the speaker does not know that the team will certainly, or most probably, win the match – for instance because the match has been fixed. In the Prior Analytics (I, 9–12) Aristotle also addressed mediated inferences involving modalities and tries to develop a modal syllogistics, that is a “theory of syllogisms from premisses some or all of which are apodeictic (necessary) or prob- lematic (possible) as distinguished from plain (assertoric) premisses” (Hintikka 1973: 135). An example of modal syllogism, derived from Prior Analytics (I, 15), is the following, which Aristotle considered valid (cf. Hintikka 1973: 137): (14) (14.a) It is possible that all female spectators in the theater like the movie. (A possibly applies to all B) (14.b) All spectators in the first row are female. (B applies to all C) (14.c) It is possible that all spectators in the first row like the movie (A possibly applies to all C) Contrary to the immediate inferences of the square, Aristotle’s modal syllogistic was not to become a stable acquisition for logic. The exact interpretation of Aristotle’s modal syllogisms remains to this day obscure and the claim to validity of several syllogistic forms presented, including the one illustrated in (1),20 is regarded as questionable by modern logicians (cf. also Rescher 2006b). In the Middle Ages important developments in modal logic were fueled by the needs of metaphysical and properly theological speculation. The very idea that we should recognize different sorts of modalities rather than just one kind of possibility and necessity emerges with the medieval thinkers. On the one hand, Medieval philosophers include in their treatment of modalities a variety of predicates expressing knowledge, doubt, opinion, appearance, volition and agency (cf. Uckelman 2012; Rocci 2005a: 79; Colmegna 1984: 174–178) some- what anticipating the modern conception on non-alethic modalities (such as the deontic concepts of obligatoriness and permissibility used in the examples above). 20 The problem with (14) lies in premise (14.b): the conclusion (14.c) is validly deduced only if (14.b) holds true in all relevant alternatives rather than simply happening to be the case contin- gently. In other words, (13.b) needs to be covertly understood as necessary in some relevant sense. “Otherwise – as observed by Hintikka (1973: 137)— it might happen that then the Bs to which A possibly applies are not Cs any more when this possibility is actualized”. In term of the movie theater example: suppose that the possibility that all female spectators like the movie is indeed realized only tomorrow: the fact that today all first row spectators happen to be female does not tell us anything about the occupation of the first row tomorrow. To infer (14.c) we would instead need to know that the seat of the first row are always reserved to female spectators. 22 1 Introduction On the other hand, when it comes to alethic possibility and necessity, they are brought by their philosophical and theological needs to broaden the range of possi- bility. While for Aristotle, a possibility is, has been or will be actualized at some moment in time (Hintikka 1973: 95), the philosophers of the High Middle Ages come to conceive of possibilities that never become the case but are still envisaged by God’s omniscience and could become the case God willing. John Duns Scotus (1266–1308) comes to conceive the contingent (bi-lateral possible) not as some- thing that is not always or necessarily the case, but as something whose opposite could be actual at the very moment when it occurs, thus introducing an idea of syn- chronic alternatives that anticipates the modern notion of possible worlds (cf. Knuuttila 2013, 1982). Scotus calls this broader, more abstract, notion of possibility possibilitas logica or potentia logica. For Scotus (cf. Vos 2006: 175), logical pos- sibility is a quality (modus) of the composition of terms (subject and predicate) which are not incompatible with each other (non sibi invicem repugnant) and can be meaningfully joined to form a proposition. Such a possibility, Scotus explains, does not require any reality apart from the semantic compatibility of the terms and should be contrasted with real possibility (potentia realis), which takes into account the structure of the natural world. In recent years, the sophisticated contributions of the medieval scholars to the field have been increasingly the object of investigation by logicians re-examining them under the lens of contemporary formal logic (cf. Uckelman 2009a). Also in this book, on occasion, it will be useful to revisit Medieval authors’ insights regard- ing certain key distinctions in the domain of modality. Indeed, as for several areas of philosophical logic, modal logic was almost re- born well into the Twentieth Century, after long period of stagnation and disregard, as part of the new – mathematically informed – formal logic. During the first part of the century, many formal logicians followed Russell’s example in regarding modal notions as leftovers of a traditional metaphysics which had no place in mathematical philosophy (cf. Rescher 2006c ). Starting with the late 1940s, however, modal logic witnessed a rapid growth thanks to the huge impulse of the possible world semantics developed by logicians such as Saul Kripke (1963) and Jaakko Hintikka (1962). Modal logic is now a huge field, with wide ranging applications, including the formalization of different aspects of argumentation in ordinary discourse. While in this book I draw from possible world semantics in Chap. 4 in order to provide a semantic analysis of modal expressions in natural language, the book does not directly contribute to modal logic. Most importantly, the present study will not deal, except marginally, with modal logic’s concern for determining which argu- ments containing modal propositions are deductively valid. As we have seen in the previous sections, the present investigation is concerned not with modality in the content of certain arguments, but rather about what modals can say about the essential structure of arguments and what role or part modality might have in such a structure. Yet, as it will emerge, over the course of the investi- gation, such a clear-cut distinction between modality in the content and modality as part of the structure can become at times quite difficult to draw. 1.6 What This Book Is Not About

Modality in Argumentation PDF

Document Details

Tags

Related

Summary

Full Transcript

Upgrade to continue