Evaluating the Language of Argument [1st ed.] 9783030616939, 9783030616946

Citation preview

Argumentation Library

Martin Hinton

Evaluating the Language of Argument

Argumentation Library Volume 37

Series Editor Frans H. van Eemeren, University of Amsterdam, Amsterdam, The Netherlands Editorial Board Fernando Leal Carretero, University of Guadalajara, Guadalajara, Jalisco, Mexico Maurice A Finocchiaro, Department of Philosophy, University of Nevada at Las Vegas, Las Vegas, NV, USA Bart Garssen, Faculty of Humanities, TAR, University of Amsterdam, Amsterdam, Noord-Holland, The Netherlands Sally Jackson, Communication, University of Illinois Urbana-Champaign, Urbana, IL, USA Wu Peng, School of Foreign Languages, Jiangsu University, Zhenjiang, China Sara Rubinelli, University of Luzern, Nottwil, Luzern, Switzerland Takeshi Suzuki, School of Information and Communication, Meiji University, Chiyoda-ku, Tokyo, Japan Cristián Santibañez Yañez, Faculdad de Psicologia, University of Concepción, Concepción, Chile David Zarefsky, School of Communication, Northwestern University, Evanston, IL, USA Sara Greco, IALS, Università della Svizzera Italiana, Lugano, Ticino, Switzerland

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. This book series is indexed in SCOPUS. 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

Martin Hinton

Evaluating the Language of Argument

Martin Hinton University of Łód´z Łód´z, Poland

ISSN 1566-7650 ISSN 2215-1907 (electronic) Argumentation Library ISBN 978-3-030-61693-9 ISBN 978-3-030-61694-6 (eBook) https://doi.org/10.1007/978-3-030-61694-6 © Springer Nature Switzerland AG 2021 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

For Agnieszka

Introduction

Only connect the prose and the passion, and both will be exalted E. M. Forster, Howards End

The purpose of this book is to seek out and examine the connections which exist between the arguments we make and the language with which we make them; to reunite the disjoined meanings of the logos, of speech and of reason, so that both may be exalted, and yet that both may be judged. In undertaking this task, my aim is not to survey and discuss merely, but to provide a subtle, functioning, tool with which to analyse and assess: all that comes before in these pages is a preparation; a clearing, levelling, and hardening of the ground, for a scheme of analysis based upon a semantics of informal argumentation. That semantics in turn rests upon an understanding of what language is and how it comes to bear meaning, of what argument is and how it comes to be expressed, and of what errors in argument consist in and how they may be recognised, exposed, and, perhaps, corrected. The project of which this work is the culmination began as a simple idea to reexamine the so-called fallacies of language, born out of some dissatisfaction with their treatment in the literature on fallacy theory. As is generally the case with ideas worth pursuing, this one expanded at a terrifying rate: to improve on the descriptions of linguistic fallacies already in existence, mine should be more thorough, more systematic, better supported by a theoretical framework, and, crucially, a useful aid rather than an esoteric curiosity. It should, in short, be grounded in a clear and solid theory of argument, and lead to an applicable semantics of argument in natural language. These requirements necessitated some engagement with theories of language, and specifically with theories of meaning. The results of this engagement are set out in Part I of this book, and they will take us some way from the safe waters of argumentation, through the treacherous straits of Thought Itself and into the perilous seas around the Cape of False Hope known as the Philosophy of Language, before returning, via Linguistics, into more familiar climes. This journey will establish two keystones of a framework for a semantics of argumentation: that language has begotten the reasoning it can express, and that language carries meaning packed with vii

viii

Introduction

that reasoning. In other words, that language has created our ability to argue, and that language is saturated with the elements which go into it: it would not be possible to argue without speech, and it is not possible to speak without arguing. In order to properly describe fallacies of language there is a need for a theory of fallacy; but in order to explain how arguments go wrong, there is a need for a theory of argument. Part II lays out such a theory. It includes a definition of what argument is, or rather how it is to be understood here, and a description of a number of modes of argument. One might question whether it was quite necessary to construct one’s own theory given that others have done so in the past. These chapters do contain a good deal of discussion of argument and argumentation from the leading scholars of the last half-century or so, and the conclusions reached do lean heavily on the insights of their work. Theories start from definitions, however, and definitions, as will be seen, are tailor made for the purposes of the theories that will wear them. My own is no different: I do not attempt to define, describe and delineate what argument is, for readers will already be well aware of the meaning of the word and the practice of the activity, rather I set out what I shall take argument to be. The principal structure of my description of argument is a tripartite division into Reasoning, Expression and Process. I shall make arguments of my own to commend this approach and seek favour for it by showing that it is a reasonable and fruitful way to proceed, nestling comfortably into the wider networks of informal logic and discourse linguistics which make up argumentation theory. With these preliminaries completed, and a theory of what argument is in place, it is possible in Part III to examine errors in argument. The three chapters deal with fallacy in general, how previous descriptions fit with the conception of argument being used herein and how that three-part framework allows for a more orderly categorisation of fallacies than has generally been achieved; with fallacies of language, of what it means for a fallacy to be of language, and how expression and reasoning are entwined as much in their error as in their birth; and, finally, with fallacies of language found in the practice of philosophy. It is a repeated theme throughout this work that, while errors in the expression of arguments may sometimes be egregious and easily identified by a critical thinker, they are, in fact, often errors of the subtlest kind; that they may hide from the finest of minds engaged in the finest of distinctions and that, therefore, they are a particular plague among philosophers, even to the point of being the cause of much of the debate within that field. The final part of this book is the fulfilment of its purpose: I set out a scheme for the assessment of arguments based on the theoretical structures put into place by the earlier chapters. The general scheme includes all three elements of argumentation, though much is left to be filled in at other times or by other authors at the levels of Reasoning and Process. The level of Expression is the focus of this work and here there is great detail and careful consideration both of the theoretical aspects and the practicalities of employing them in the actual analysis of arguments. The final chapter describes in detail the analysis of five examples by use of the assessment scheme, illustrating how it is capable of identifying subtle errors of expression; or rather, how it can aid the analyst to do so, for this is not an automated system, and, ultimately,

Introduction

ix

arguments are to be accepted or rejected by their audience, not by a framework or set of rules. The theoretical sweep of this short volume is large, meaning much has been left out and much has been passed through at speed. There has been little room for the consideration of the classical authors, and nowhere near enough for a full survey of the vibrant and fast-flowing stream of current argumentation research. At each step I have tried to include enough theory to allow me to reach the next step, rather than to give a complete record. My purpose has been to reach the goal of a semantics of informal argument and a usable scheme for its assessment. No doubt this semantics could be developed further, and certainly the levels of analysis of Reasoning and Process should be redrawn with greater detail and finer skill. As we go forward, and even as we look to divide up and define, I repeat the exhortation with which I began: Only connect! For a connection is nothing but a leap, an inference, an argument.

Contents

Part I

Language

1

Language and Thought . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3

2

The Meaning of Meaning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

17

3

Language as Argument . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

31

Part II

Argument

4

What Argument Is . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

45

5

The Modes of Argumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

59

6

Pragmatics, Rhetoric and Semantics . . . . . . . . . . . . . . . . . . . . . . . . . . . .

79

Part III Error 7

Fallacies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

97

8

Fallacies of Language . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

9

Linguistic Fallacies in Philosophy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

Part IV Analysis 10 Analysis of Arguments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 11 CAPNA—The Comprehensive Assessment Procedure for Natural Argumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 12 Application of the Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 xi

List of Figures

Fig. 11.1 Fig. 11.2 Fig. 11.3 Fig. 11.4 Fig. 11.5

CAPNA. A comprehensive assessment procedure for natural argumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A scheme for the analysis of argument process . . . . . . . . . . . . . . A scheme for the analysis of argument reasoning . . . . . . . . . . . . . IAS. An Informal Argument Semantics . . . . . . . . . . . . . . . . . . . . A scheme for the deep linguistic analysis of arguments . . . . . . . .

169 173 177 181 188

xiii

List of Tables

Table 10.1 Table 11.1

A partial typology of fallacies . . . . . . . . . . . . . . . . . . . . . . . . . . . Suggested procedural questions for Informal Argument Semantics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

155 193

xv

Part I

Language

Words are not just wind. Chuang Tzu

Chapter 1

Language and Thought

In this first chapter, I want to examine the fascinating and interdependent relationship of thought to language, and begin the process of describing how both relate to argument. I am going to pursue the claim that just as language must be the result of thought, so too thought, understood as a process of reflection to which the human mind has access, is, in some sense, created by language; that languageless thought is mere instinct and emotion, and higher cognition is only possible through the linguistic conceptualisation of experience. Understood in this way, all thought is reasoning, and all expressed thought is argument. Not everything which we say actually expresses a full argument, of course; but every utterance is the result of a process of reasoning. That means that every utterance is, on some level, susceptible of an argumentational analysis. This is true not only of statements, which can be understood as conclusions drawn from the available data, but also of questions and imperatives. Issuing an order is no different from concluding that the subject of the order ought to perform a certain action in order that some end be achieved; asking a question is to reach the conclusion that certain information is lacking which is necessary to one’s reasoning: one has concluded that such information is necessary. The utterance then may be either an expression of the reasoning or an enactment of its conclusion, from which the argument may be inferred. It is a common practice, and one which we shall have further cause to discuss, for theorists to reconstruct argument forms from the logically incomplete structures of actual discourse. Theoretically, this could be done for every utterance; though it is possible to play all sorts of games with language—we can ask questions just to see if someone knows the answer even when we already possess it, we can write poems and stories, we can elaborately express emotional states—still, every utterance is the result of a reasoning process which has concluded that that particular utterance is the one to be produced at this moment: any other sounds we produce are just noise. The first part of this book is dedicated to an analysis of the nature of language, and that must begin with a consideration of the nature of its creator. Language is what divides us from the animal kingdom; it defines our humanity. Language is a product of the lived human experience; without the context of that experience it cannot be understood and it could not have existed. It is not a system for describing the world, © Springer Nature Switzerland AG 2021 M. Hinton, Evaluating the Language of Argument, Argumentation Library 37, https://doi.org/10.1007/978-3-030-61694-6_1

3

4

1 Language and Thought

it is a system for describing the human experience of the world: it is not a system for codifying the truth, but a system for expressing human perceptions of the truth. Words reflect thoughts, not reality. As the chapters of this work progress, we shall see how an over-estimation of, or better, a misplaced reliance upon, the qualities of language, leads humans, particularly those who would think deeply and must, as humans, pursue their thoughts through the medium of words, into muddle and confusion—what Wittgenstein referred to as ‘philosophical bewilderment’ (1958: 1).

1.1 The Origin of Language In this section, I begin from the very beginning in an attempt to understand what the nature of language is and how it affects the phenomenon of argument. I briefly discuss some theories of the origin of language, but rather than trying to reach a conclusion as to the historical course of events, I shall use the debate around them as a means to draw out certain aspects of language, of what it is, what is it for, and, crucially, how it relates to thought—since I suppose that everyone would agree that without thought, there could be no argument. I shall not dwell on the physical aspects of speech, on the evolution of the speech organs, or even the brain; not because those questions are not interesting, but because they bear no relevance to the tasks ahead. A great many animals, and certainly the familiar mammals, make sounds of one sort or another: there is nothing unusual about the human capacity to do so. We may be sure that without possession of language, humans would not be silent: there would be roaring and snarling, screaming with pain and frustration, grunting and groaning, perhaps even sighing and laughing. It is not the case either that the ability to produce the range of sounds familiar to us as human speech is a recent development in our evolutionary history: in their review of the fossil record, which takes into account skull shape, respiratory systems, oral cavities and bones in the throat, Frayer and Nicolay conclude that: ‘in homo dated to around 1.5 million years ago, a set of morphological features emerged that are correlated to speech sound production in modern humans’ (2000: 231). This suggests that the biological capacity for the production of speech, and indeed song, existed in our ancestors for a million years or more before the emergence of the modern human. Having the capacity to make the right noises and having the brain power to do so meaningfully are, of course, two very different things. This makes life difficult for anyone wanting to identify when words were first used. Not only is there no way of knowing what sounds were actually being made, or for what purpose, it isn’t at all clear how one would draw the line between words and other noises. In modern linguistics, interjections are classed as words, but just how conventionalised does an expression of pain have to be for it to be considered a word? Some of the theories for how the first words came to be spoken are quite good fun, and tend to be based on a creative imagining of the lives of the first speakers. In his

1.1 The Origin of Language

5

classic list of 1861, Max Müller (1864), includes the theory of words developing from interjections; also as imitations of the sounds of the natural world; as well as what is known as the ‘yo-heave-ho’ theory, by which the rhythmic sounds of coordinated labour led to distinct utterances such as ‘heave’. A more modern suggestion is that of Dean Falk (2004) who describes how the necessity for human mothers of putting babies down while foraging, rather than having them cling to their fur, as other primates do, meant that caring mothers—those who children were more likely to survive—used sounds to reassure and encourage their children when physical contact was not possible. This theory, in keeping with the modern trend, addresses the social and emotional conditions necessary for language to begin, something in which earlier theories showed less interest; a similar principle has been applied by Oren Poliva (2015, 2016). There are a couple of advantages to theories which involve babies: firstly, they explain how fairly basic sound making ability could have helped in terms of survival, since babies die easily, and secondly, they give a reason for individuals to be exposed to what language input there was from a very early age, just like modern humans; an exposure which we know is necessary to develop the highest abilities in any given language. Still, there are reasons to be doubtful as to how far this sort of communication could have gone. Even when exposed to relentless all-day chattering, infants learn very little language before they learn to walk. The window in a child’s development between becoming aware of the meaning of its mother’s calling and being capable of following her around on its own two feet would have been rather short, certainly too short to have built up much of an ability for repartee.

1.2 From Words to Language However humans came to utter the first words, we know that they did. Those words were not a language in the sense that we understand today. Language had to be developed: it must have developed for a reason, and it must have developed in the way it did for a reason. Language is simply too complex and too biologically expensive, in terms of the required brain power, to have appeared, as it were, by accident. This is certainly a question with plenty of room for speculation: the practical man will assert that men learnt to communicate in order to solve practical problems or to better coordinate their efforts when working together; the artist may suggest that it was in pursuit of the expression of the poetic that man first formed structures from his words; the religious see language as a gift from God which allowed men to praise Him as well as to criticise each other. I would suggest a motivation more in keeping with our instincts: in the animal kingdom it is for the rituals of mating that the most sophisticated, exaggerated, and flamboyant behaviours are reserved, and it is my belief that for the human race too, language grew out of the desire to find and keep a mate. For man is an industrious animal, and he is a practical animal, and yet, far more, he is a lustful one. Not for him the confining of mating to one season of the year,

6

1 Language and Thought

nor even to the process of procreation. It is man’s lust that drives him on; it is his lust that makes him strain, for he strains not for the results of his efforts, but for the glory they cast upon him amongst the group in which he lives. Since in all else man is driven by his lust, why not then in the founding of his language and the beginnings of his dialogue? Charles Darwin himself wrote: ‘A strong case can be made that the vocal organs were primarily used and perfected in relation to the propagation of the species’ (Darwin 1888: 566), and Wolfgang Wildgen explains: ‘Even if linguistic communication is not a basic precondition for sexual partnership, it is rather natural to assume that the sexes (one or both) developed a preference for communicatively proficient partners and selected in majority on this criterion’ (Wildgen 2004: 19). In all likelihood, we shall never know what drove words forwards into language, it may have been a great many different things, and perhaps it doesn’t much matter. Two points are worth making in the context of this study though: firstly, if we conceive of language as having grown out of personal, intimate communication, we can understand it as a means of expressing the human experience, rather than as describing or attempting to shape the world, a view which, I think, fits better with modern linguistics. Secondly, given the strong motivations, on both sides, to beget children and to have them properly looked after, and considering the huge risk taken by the female, it is natural that skills of persuasion, explanation, reassurance, negotiation and bargaining would have developed: argumentative discourse is the discourse, everything else is gossip and entertainment.

1.3 The Relationship of Language and Thought There are two main questions to be addressed and perhaps answered in this section: language is clearly built by thought, but to what extent does language affect that thought? And, before that, what types of thought are necessary to make language possible? Since both of these questions represent research fields on their own, what follows will not be an attempt to summarise all the various theories in circulation, but rather to note certain points of interest having bearing on argumentation theory and influencing the themes of the later chapters of this work. It will also be seen that keeping the two questions apart is a very difficult task, such is the entwined nature of the two phenomena. There is a danger of sinking into a chicken and the egg scenario when considering the interdependence of language and thought. Certain thoughts, or types of thought, may seem to be impossible without language, and yet language can hardly have been developed by individuals who had no thoughts. This leads towards some very slippery places indeed, as it suggests that the way forward is to define thought, and possibly to distinguish thought from thoughts, a path I have no desire to tread at this moment. Descartes, in the second meditation, famously described himself as ‘a thing that thinks’ which he then defined as ‘a thing that doubts, perceives, affirms, denies, wills, does not will, that imagines also, and which feels’ (Descartes 1968:

1.3 The Relationship of Language and Thought

7

107) which, while no doubt meant to offer some explanation, is a perfectly useless list of synonyms telling us nothing at all about what it means to think. What can be said with some certainty though is this: for language to have got started, humans must have had some theory of other minds, they must have felt that whatever was in their minds could be communicated to, and understood by, other humans; and for language to develop, they must have had a degree of trust in other individuals: one cannot get knowledge from language, either about the world or about the speaker, if one does not believe what one hears. In a world where everyone is suspected of lying, what would be the point of talking? Wildgen in describing the development of linguistically skilled groups combines these two points: In general the new quality of social organization required a kind of “theory of mind”, which allowed partners to trust one another (some authors say romantically that sex changed into love under the new conditions). Information about others’ attitudes, beliefs, and dispositions to act can only be transferred if a language beyond the expression of actual feelings and reactions to present situations exists. (Wildgen 2004: 20)

This requires some giant mental steps. If we are to gain or even seek information about others’ beliefs, we must first have some idea of ‘belief’. One of the best known theorists to have described language as essential to thought is Donald Davidson, who wrote: ‘a creature must be a member of a speech community if it is to have the concept of belief. And given the dependence of other attitudes on belief, we can say more generally that only a creature that can interpret speech can have the concept of a thought’ (1984: 170). The only way one can have beliefs is to recognise that the way the world appears to be to one’s own senses is a fallible construct; that others could see things differently, and that there is information outside one’s own knowledge which could alter one’s understanding of the world: in short, to recognise the concepts of truth and error. For Davidson, this realisation comes through the process of interpretation: it is the interpretation of the speech of another which brings about a grasp of his beliefs, ‘if we know he holds a sentence to be true and we know how to interpret it, then we can make a correct attribution of belief’ (his italics) (1984: 162). Beliefs, then, seem to be what we can infer about the mental states of others from what they say. ‘For the notion of a true belief depends upon the notion of a true utterance, and this in turn there cannot be without shared language’ (1984: 70). This would appear to bring us back to the point of asking what is meant by ‘thought’. Davidson suggests that all attitudes depend on the notion of belief, and it seems that by thoughts he means what are also known as propositional attitudes, which bring with them the notion of intentionality (Pérez 2005; Jacob 2019). The suggestion is not then that the mind is incapable of any activity without language, which would make the emergence of language impossible, but that thoughts which are intentional, thoughts about something, only develop once language interpretation is undertaken.1 1 The

suggestion that there is a language of thought (LOT) which is prior to language acquisition is less interesting than it sounds. What is meant by that is that primal thought has structure and therefore syntax: ‘linguistic capacities are systematic, and that’s because sentences have constituent structure.

8

1 Language and Thought

Although Davidson’s theory of ‘radical interpretation’ of which all this is part, is controversial, we are now moving in the right direction: the point to be made is that some types of thought are only available to beings with language. This has been recognised by many thinkers. David Kaplan wrote: ‘On my view, our connection with a linguistic community in which names and other meaning-bearing elements are passed down to us enables us to entertain thoughts through the language that would not otherwise be accessible to us’ (Kaplan 1989: 602); before that Frege: ‘without symbols we would further hardly raise ourselves to the level of conceptual thought […] since the concept is of itself imperceptible to the sense, it requires a perceptible representative in order to appear to us’ (Frege 1882: 156). We are now drifting into the arena of our second question, how language affects thought, but, finally, we come to an intersection with the theory of argumentation. The ability to argue is clearly dependent on the possession of language, since arguing is a linguistic activity, but it appears to be the case that the ability to think argumentatively is also dependent on language. Language is not only a tool to express our reasoning, it is a precondition of that reasoning. That doesn’t mean that all thought is reducible to language, even Davidson says of his theory ‘it allows that there may be thought for which the speaker cannot find words, or for which there are no words’ (1984: 158), rather that the ways in which we think, the ways in which we form ideas and make connections are influenced by our knowledge of language, and are indeed, a result of it. Davidson explicitly draws the connection between interpreting behaviour and the exercise of reason: The cogency of teleological explanation rests, as remarked, on its ability to discover a coherent pattern of behaviour of an agent. Coherence here includes the idea of rationality both in the sense that the action to be explained must be reasonable in the light of the assigned desires and beliefs, but also in the sense that the assigned desires and beliefs must fit with one another. (1984: 159)

We can give an explanation of another’s actions by arguing that the desires and beliefs we assign to him, our premises, are both consistent and lead rationality to his behaviour, our conclusion. Indeed, following this line of thought, we might suggest that rationality itself is a product of language. We may assess the behaviour of an animal or an infant as ‘irrational’ in our terms, but we would hardly say that the animal was being irrational, since it doesn’t have the capacity to reason. And could a non-linguist be incoherent? Coherence and consistency are properties of beliefs or assertions, not fleeting impressions in the brain, so presumably not. The key properties of good argument then, rationality, coherence, cogency, are all products of the existence of language, they are words with no meaning when applied to non-verbal environments.

But cognitive capacities are systematic too, and that must be because thoughts have constituent structure. But if thoughts have constituent structure, then LOT is true’ (Fodor 1989). Such debates are of consequence to philosophers of mind and cognitive scientists, but have no bearing on the issues here.

1.3 The Relationship of Language and Thought

9

At the same time, other kinds of non-intentional brain activity are obviously present both before language learning in humans and in other animals. The difficult question is how the non-linguistic gave rise to the linguistic. As was noted above, trust seems to be necessary, but trust would be developed through speech which turned out to be true; and a sense of the minds of others is necessary, but that too would come about as a result of their expression of their thoughts. Fortunately, we need not concern ourselves with this conundrum, except to say that the complexity of language and the sophistication of human thought must have developed hand-in-hand, so to speak, with each step in the one helping the advancement of the other. Language, then, starts out simple, as it must, but the human brain, then as now, has massive capacity for symbolic interpretation. That potential is released by exposure to language, but it is likely that the development of what we know as thought leads to a weakening of other areas of brain activity, the sensory and instinctive: while we as a species have undoubtedly gained a great deal through our development of language, it is impossible to know what we have lost. The development of the cognitive skills associated with language: the ascription of beliefs, abstract reasoning, narrative memories, personal identity, and more, runs concurrently with the development of language skills and occurs anew with each individual. How rapidly, and how often, language evolved from the first words with symbolic meaning to a complex system which affects every aspect of our mental life we may never know; but at some point, it becomes possible to understand that one holds a belief which is different from that held by another and to offer reasons in its support, and at that moment argument is born.

1.4 Sapir-Whorf and Linguistic Relativity In this well-known debate, John Lucy provides some important distinctions, separating out three levels of linguistic relativity: Semiotic, where the presence of any symbolic system affects thought; Structural, where knowledge of a particular natural language may have a certain influence; and Discursive, where certain language practices affect thinking (Lucy 1997: 292). The issues we have looked at so far have fallen into the category of the Semiotic, they have dealt with the way that the brain interacts with the phenomenon of language, with the incredible development of symbolic meaning. Linguistic relativity, however, has more often been taken to mean what Lucy calls the Structural. The so-called Sapir-Whorf hypothesis is, at root, the claim that languages, rather than just language, affect thinking, and so those who speak different languages think differently. Sapir was a student of Boas, who had worked to show the complexity of unwritten, non-European languages, thus weakening the theory of the evolution of languages prevalent in the nineteenth century and somewhat similar to the racial theories of the time. This work naturally led towards a view of each language representing a unique way of conceptualising the world, rather than all languages being inferior versions of the great literary languages of Europe; and that would mean

10

1 Language and Thought

that those who knew only those languages would know only that conceptualisation. Whorf was a student of Sapir and it is his writings which form the basis of the theory bearing both their names. The extent to which it is a theory in any scientific sense is disputed: Paul Kay and Willet Kempton note that ‘there have been many attempts to recast the fundamental insights of Sapir and Whorf, originally expressed in a number of evocative and sometimes metaphorical passages’ (Kay and Kempton 1984: 66), and Barbara Scholz et al. (2016: §4) are even less complimentary, calling their writing ‘more often polemical than enlightening. It certainly holds no general answer to what science has discovered about the influences of language on thought’, denying that Whorf offers any hypothesis at all. What Whorf did say was this: Formulation of ideas is not an independent process, strictly rational in the old sense, but is part of a particular grammar, and differs, from slightly to greatly, between different grammars. We dissect nature along lines laid down by our native languages. The categories and types that we isolate from the world of phenomena we do not find there because they stare every observer in the face; on the contrary, the world is presented in a kaleidoscopic flux of impressions which has to be organized by our minds—and this means largely by the linguistic systems in our minds. We cut nature up, organize it into concepts, and ascribe significances as we do, largely because we are parties to an agreement to organize it in this way—an agreement that holds throughout our speech community and is codified in the patterns of our language. The agreement is, of course, an implicit and unstated one, but ITS TERMS ARE ABSOLUTELY OBLIGATORY; we cannot talk at all except by subscribing to the organization and classification of data which the agreement decrees. (Whorf 1956: 212–214, emphasis in original)

Which is tidied up and divided into two separate hypotheses by Roger Brown: I. Structural differences between language systems will, in general, be paralleled by nonlinguistic cognitive differences, of an unspecified sort, in the native speakers of the two languages. II. The structure of anyone’s native language strongly influences or fully determines the world-view he will acquire as he learns the language. (Brown 1976: 128) The disjunction in the second hypothesis ‘influences’ or ‘determines’ illustrates how Whorfism can be divided into weak and strong forms, a distinction made popular by Julia Penn (1972) among others. The weak version is also known as, or linked to, linguistic relativism, and the strong as linguistic determinism, and in both forms it has been largely dismissed by linguists. In their review of experimental work done at that time, Kay and Kempton conclude that: ‘languages differ semantically but not without constraint, and second, that linguistic differences may induce nonlinguistic cognitive differences but not so absolutely that universal cognitive processes cannot be recovered under appropriate contextual conditions’ (1984: 77), or as Scholz et al. put it: ‘The weak versions are commonly dismissed as banal (because of course there must be some influence), and the stronger versions as implausible’ (Scholz et al. 2016: §4.4). The reasons that stronger versions of the idea are not considered plausible are manifold; but most obviously while languages are structured differently and use grammar to express what needs to be said in various ways, it is not particularly

1.4 Sapir-Whorf and Linguistic Relativity

11

problematic for speakers of other languages to see the relevant distinctions and to learn the other language. Much of the research on Sapir-Whorf has, in fact, focussed on lexical rather than structural differences. In the passage cited above Whorf claimed that speakers divide nature into concepts, which is a common enough assumption, and it is clearly the case that some languages have words for which others do not have an equivalent, and some divide the scope of a word in one language into many words in their own. The linking of words and concepts can be problematic, however. The presence of two words does not prove the existence of two entities and the absence of a word does not mean that the idea is unknown to speakers of the language. The arbitrariness of the way words and ‘concepts’ combine can also be shown in the English borrowing from German, ‘schadenfreude’. The fact that one word to describe the pleasure one takes in the misfortune of another does not exist in English, does not mean that the Germans have had this concept for generations and the English only later discovered it, any more than that the concept of taking pleasure from inflicting pain was created when de Sade’s name was given to Sadism. One word may cover multiple uses, evoking apparently different ‘concepts’ and one ‘concept’ may require multiple words to express it, but there is nothing special about the meanings which need only one word to express, or anything unified about the meanings which can be expressed by the same word. That said, however, it seems reasonable to suppose that generally, in both private thought and spoken language, we reach most readily for that which is close at hand, and there may be combinations of ideas which have not previously occurred to us, or whose expression seemed unnatural and awkward, which are easily evoked in an elegant new word. Arne Naess argues for this view: ‘a moment’s reflection makes it clear that the direction and scope of all our thinking is largely determined precisely by what short concept terms are available. […] the introduction of new concept terms opens the way to new possibilities for thought’ (1966: 68), which leaves a delicate uncertainty as to whether in the new concept term the key element is the concept or the term, or whether indeed it makes sense to try to separate them. Despite the doubts, it ought to be acknowledged that the Sapir-Whorf hypothesis has never gone away, and there has been a constant flow of research into linguistic relativism. In their 2015 paper Jordan Zlatev and Johan Blomberg consider several aspects of importance to any empirical work on the language/thought connection. Firstly, they seek to show that language and thought should be properly disentangled and suggest a definition of language where one of its purposes, along with sharing with others, is ‘enhancing cognition’ (2015: 2) a process which allows us to reach the higher cognitive functions. It is this ‘mediated cognition’ which is the ‘thought’ which is affected by language, and it is the extent and the form of that mediation which is under investigation when we study the effects of language on thought. They also acknowledge the role of non-linguistic cultural factors and the need to dissociate them from linguistic factors, a theme taken up by both Enfield and Kienpointner, as described below. Thirdly, they stress the need to recognise that ‘different kinds of (possible) linguistic influence on thought need to be distinguished, and some may be more pervasive than others’ (2015: 6).

12

1 Language and Thought

N.J. Enfield (2015a) provides an excellent review of both empirical and theoretical work, which suggests that the above considerations have been well understood across the field. He also gives a view of the importance of the topic, which revolves around decision-making heuristics, and conceptualisation: Concepts are tools in decision making because they give us criteria for recognizing instances of what we are looking for and of what we are not looking for. In this way, concepts are sieves (Kockelman 2013; Enfield 2015[b], p. 172). They are means for deciding, and thus for acting. Whorf’s idea was simply that language is likely to play a central role in decision making along these lines. At the root of this idea are two facts: (a) the most ready-at-hand concepts are the ones encoded in the languages we speak, and (b) the languages we speak are known to differ radically in the kinds of concepts they encode. This suggests a prima facie argument for a form of cultural relativity grounded in differences between languages. If concepts provide a basis for categorization and decision making, and if different languages supply their speakers with different concepts, then different languages provide their speakers with different bases for decision making and, subsequently, for different patterns of behavior. (Enfield 2015a: 210)

The emphasis here on decision making naturally leads the argumentation scholar to think of practical reasoning. Even if we think of logical deduction as fixed and independent of language differences (which I’m not sure we should), it is interesting to consider how abductive, defeasible reasoning may be influenced by the categorisations made in the language we speak. Any inference to the best explanation could easily be at the mercy of such conceptualisations. This would make an interesting, if challenging, subject for study. One of the studies Enfield mentions in the newer wave of research is that of Wolff and Holmes (2011), who present a list of five ways in which the effect of language on thought may be visible: Thinking for speaking, where the way we think is guided by the way we plan to express our experience later on; Language as meddler, where linguistic representations clash with non-linguistic ones; Language as augmenter, where linguistic representations enhance non-linguistic ones; Language as spotlight, where language increases the salience of aspects of non-linguistic thinking; and Language as inducer, where certain non-linguistic processes are primed by language. This shows both the breadth of possibilities for research and the way that the original Whorfian ideas have been extended: and Enfield believes, on the basis of the numerous studies cited, that ‘we see many ways that language diversity correlates with patterns of cognitive diversity’ (Enfield 2015a: 213). Enfield’s own work focuses particularly on the linguistic function of building a social reality, and the types of actions which can be performed with language: ‘Only when we frame language in terms of agency will we properly understand the ways in which it is used to create social reality’ since ‘users acquire, claim, and develop rights and duties in relation to the defining elements of linguistic agency: flexibility and accountability’ (2015a: 218). On the basis of his review of recent research, Enfield concludes that: ‘Linguistic diversity is causally related to cultural and cognitive diversity in numerous ways. Some effects of language on thought are now widely acknowledged to be real. But their possible interactions with the true diversity of human languages remain

1.4 Sapir-Whorf and Linguistic Relativity

13

understudied’ (2015a: 219). What this cognitive diversity means for both crosscultural argumentation and argumentation viewed cross-culturally, it is difficult to say.

1.5 Language, World View and Argument The title of this, the final section in this opening chapter, is borrowed from an essay by Manfred Kienpointner and will contain a discussion of his attempt to draw conclusions from some of the issues which have been discussed above in respect of their impact on argumentation theory. One of the apparent obstacles to research on the influence of language on thought pointed out by both Enfield and Zlatev and Blomberg was that it is often based on a misconstrued and overly narrow conception of language. Kienpointner takes an overtly Wittgensteinian language-game approach, which straddles the types of relativity referred to by Lucy as Structural and Discursive: the structure of a particular language cannot be separated from the uses to which it is put, and those playing different games in what appears to be the same language are actually conceptualising differently. Wittgenstein wrote that ‘the term “language-game” is meant to bring into prominence the fact that the speaking of language is part of an activity, or of a form of life’ (1953: 11, §23, original emphasis). This means that it is not merely the words and grammar that we know, but also the way in which we use them which is referred to when we talk of the languages we speak. It is this combination which affects the way we see the world, not just abstract structural form: ‘The embedding of language in the context of actions and activities and, ultimately, in a form of life constitutes a framework which shapes our picture of extra-linguistic reality much more than the purely structural system of our language’ (Kienpointner 1996: 481). This can lead to practical implications in the realm of argument, since those who appear to speak the same language may actually have different norms of use of certain words. Kienpointner gives political examples such as ‘fascism’, which is currently enjoying a revival across Europe and North America at least as an insult if not as an ideology, and others could be added such as ‘populist’ and ‘liberal’, which last has different meanings on the left and right of politics and East and West of the Atlantic. When such differences are well-known there is no particular problem for debate, however, ‘specific rules of language games can form (more or less) rigid world views which are typical for subgroups of a speech community’ (Kienpointner 1996: 482), and it is these differing world views which are the greater threat to dialogue. Kienpointer quotes Lakoff’s assertion that the concepts these words denote are ‘lived by’ (Lakoff 1987: 321) and not just used in the understanding of language. Another phenomenon which may affect thinking is the ‘wisdom’ encapsulated in idioms and fixed expressions. These expressions are not part of the structure of the language and can be avoided in use, but their potential impact on how those exposed to them think on certain matters is clear. Kienpointner gives examples such as ‘boys will be boys’ and ‘time is money’: the first reinforces the ideas that boys are different

14

1 Language and Thought

from girls, that there are specific behaviours which are expected of boys, and that boys should be allowed to indulge in them; the latter suggests a commodification of time which is a rather profound and perhaps disturbing interpretation of human experience. Conventional metaphors, such as that depicting argumentation as battle, have similar effects. They work to shape our understanding of the world we live in, and unless, like linguists, we stop to analyse them, this influence may go unnoticed. Shaping what is considered the standard sense of a word or the standard way of expressing a situation may be useful for political purposes since ‘common usage […], different to the language system, very often has an ideological touch and expresses a certain world view’ (Kienpointner 1996: 483). This creates an interesting dynamic: since there is no neutral space, the struggle to change the meaning or the use of a particular word is always the struggle to replace one world view with another. In recent years, many English words have been pushed out of the publicly accepted lexicon as being insulting or disrespectful to particular groups in society. This, on the surface, is an attempt to fight back against the situation that Kienpointner describes, where standard use ‘usually favours a world view in which powerless groups and/or minorities are at a disadvantage and are discriminated against linguistically’ (1996: 483). An examination of examples, however, shows that the reality is more complicated: first, in the USA it has become unacceptable to say ‘coloured people’ but is fine to say ‘people of colour’. Now, the first term has certain historical associations with world views which segregated and valued, or didn’t value, people on the basis of their skin colour, and those promoting this norm would like such connotations to be avoided; yet the second term, while avoiding the form of language favoured by followers of racist ideologies in the past, is still a term for dividing people on the basis of the colour of the skin. Whether that division is desired or not, whether it is useful and even empowering to those to whom it refers is not for me to say, but it remains a term which enforces a world view in which skin colour is a salient distinguishing feature of human beings. A second example is the use of the acronym LGBT to describe people on the basis of their sexuality. Certainly, this is a politer expression than some which would have been used in the past, but it also carries with it an ideology, a view of the world, which is not in any sense ‘neutral’. I am not sure that transsexual people and bisexual people have terribly much in common, but the use of the LGBT term, and the extensions which are added to it in the name of inclusivity, suggests that society can be divided into two groups LGBT and not LGBT, which is an ideological conceptualisation: it both reflects the thinking of those who use it, and influences the thinking of all who are exposed to its use. All of this has an important implication for argumentation. Firstly, the kind of cultural gaps which can lead to ‘deep disagreements’ (Fogelin 2005) may occur between speakers of the same language living in the same communities, but belonging to different groups in terms of certain language games they play: not only will their use of certain terms diverge, but their understandings of the world, reflected in the rules of those games, may also be far apart. This could be a particular danger when those who believe themselves to be from roughly the same culture come to debate,

1.5 Language, World View and Argument

15

since most people are aware of cultural differences between say Europe and East Asia and are likely to be on the look-out for them. A related concern is that since speakers are usually unaware of the embedded ideology inherent in their own language, while it is possible to avoid certain expressions, we generally just use what we think of as normal: as Kienpointner notes, ‘most people do not reflect on language while using it. Therefore, they normally continue to employ the most frequently used expressions’ (1996: 485). This means that a speaker may not realise exactly what he is committed to in an argument, because he does not realise that those common expressions commit him to anything at all. For real world argument to be possible, theories, such as pragma-dialectics, accept that there must be a certain level of initial agreement, that is they ‘accept that speakers have to start out from commonly accepted assumptions which can be used as premises in a discussion’; this, however leads to the danger that standpoints may be considered as having been successfully defended on the basis of those initial premises, which are themselves only provisionally true: ‘This is even more dangerous if ideological assumptions appear in the premises only implicitly or if they are presented in such a way that they appear to be unquestionable’ (Kienpointner 1996: 486). In conclusion, Kienpointner stresses once again that ‘it is not language as a system, but the use of language according to the rules of language games which connects language, thought and world view’ (1996: 492, original emphasis), which is both an important clarification for the empirical study of Whorfianism and one which makes the task considerably easier. For argumentation, the considerations of this chapter have led to the following conclusions: firstly, that the types of thoughts we can have and the ways in which they can be expressed are closely intertwined, and that the very notion of reason is inextricably linked to the use of words, making the separation of the two meanings of the Greek logos, on some level at least, a mere illusion. Secondly, that conceptualisation, and different understandings of the link between a concept and the expression it is apparently represented by can lead to confusion, ambiguity, and even mysticism. Finally, that every user of language is shaped in his thinking not just by that language, but also by the use which is made of it, by the expressions which its users have normalised and the implicit ideology that they contain. A lack of awareness of this unseen force influencing our world view and the assumptions which we take into any argument as foundational premises is likely to lead us into prejudice and error.

Chapter 2

The Meaning of Meaning

That we must address the issue of what it means for something to mean something else shows in itself that the answer is not obvious. Indeed, philosophers of language spend the better part of their time on this very question. At the same time, the simplest man can use language and he knows just exactly what he means by the things that he says. In order to make some sense of this apparent paradox, it will be necessary to tread gingerly into the cool waters of the philosophy of language, whilst being careful not to get out of our depth. If there is one place in human discourse where meaning breaks down more than any other, it is in those works dedicated to building it up. Before addressing the work of some of the prominent thinkers on meaning, remembering at all times that this volume has as its goal an understanding of the relationship between language and argument, not a complete theory of language, I shall make some general remarks on the very idea of meaning and the practice of definition.

2.1 Types of Meaning There are four meanings, by which I mean uses, of the word ‘mean’ in its various forms which I can readily identify: three of which are capable of felicitous employment and one of which is incoherent, by which I mean that it is not so capable. Firstly, meaning can be used as ‘standing for’ or ‘symbolising’ (type 1). This is the sense in which words and other symbols are said to have meaning. A sign somehow comes to ‘stand for’ something else and meaning is born. When we ask, ‘what does this word mean?’ we are asking what does it symbolise. What marks out meaning of this form from other ‘representational’ signs is that the symbol is arbitrary, although the cases of hieroglyphics or pictograms are open to argument. Certainly, in this type of meaning the sign can be arbitrary. The connection—however that is understood or achieved—is the result of an agreement, in a sense tacit, yet regularly reinforced, © Springer Nature Switzerland AG 2021 M. Hinton, Evaluating the Language of Argument, Argumentation Library 37, https://doi.org/10.1007/978-3-030-61694-6_2

17

18

2 The Meaning of Meaning

by a community, of at least two speakers, that some mark or sound should be taken to stand for something else. It is not an easy thing to describe what this ‘standing for’ really is, and having spent considerable time with the literature I am not at all confident that it lies within the competence of philosophy to do so; but it is the essence of language, and the development of the ability to allow a mark or sound to stand for something else is the single greatest achievement of mankind, besides which the invention of the wheel or the planting of a flag on the moon are of the slightest significance. The second use of meaning is often employed to describe causes and effects, but can be most simply defined as ‘meaning is inference’ (type 2). When an unfortunate student is told he has failed an examination, he may ask ‘what does that mean?’, by which question he enquires about the consequences, the effects of his failure. His tutor may exploit the ambiguity and decide to deliberately misunderstand in order to make, in his opinion, a more important point: ‘It means you didn’t study properly.’ In this way, he brings up the cause, not the effect of the fact of the poor performance; and whilst he may suspect that this is not the answer the candidate was looking for, it is still a reasonable one given the phrasing of the question. Another example makes the point that the meaning of an event or circumstance can refer to all inferences which can be drawn from it. I arrive home with a friend to find the door locked, and I have no key. ‘What does that mean?’ she asks. I reply that it means everyone has gone out, that there is no-one at home and no way to get inside, and that we shall have to sit on the doorstep until someone turns up. Thus, the meaning of the circumstance may be an event in the past, an implication for the present, and a prediction of the future. In this use, then, meaning clearly refers to all the inferences which can reasonably be made on the basis of the event, circumstance, or, indeed, utterance under consideration. Third, when applied to a person, meaning generally refers to intention (type 3). Constructions such as ‘what do you mean to do?’ show that in certain phrases mean and intend can be swapped with no great consequence: intend perhaps being less common in everyday spoken English. Later chapters will warn against taking linguistic phenomena as proof of anything other than themselves, however, the evidence here is strong that this form of meaning is distinct from the others. When a communication is unsuccessful we do not ask, or even wonder to ourselves, ‘what do those words mean?’, rather ‘what did you mean (by those words)?’; and more, even when we understand the words used perfectly and their significance in the type 1 meaning above, we may still ask ‘what do you mean by that?’. Most clearly of all, after an accident, it is normal to say ‘sorry, I didn’t mean to do it’, stating that my action was not deliberate, it was not my intention. In such phrases, we cannot equate meaning precisely with type 1 or type 2, although it is quite possible that when we speak we intend for certain meanings of those types also to be understood. There is, however, a definite link which allows us to think of ‘meaning’ here as one word with different senses rather than different words with the same sound. In all three we can see that something which is apparent can be used to move beyond and discover something which is not apparent: that may be what a word stands for, it may be conclusions which can be inferred from an event, or it may be intentions

2.1 Types of Meaning

19

which are revealed in acts, including of speech. It might be argued that all moving from one thing to another in this fashion is inference. I am not going to resist that idea particularly, save to say that I think the suggestion that type 1 meanings are inferred from sounds, rather than directly understood, would require a strong argument. The point is not important anyway, since the three uses outlined are clearly distinct, and it is use with which we have most to do when discussing meaning, as will be seen below. One very important point, however, is that language may have meaning of all three types. Indeed, it cannot get by without it. That words have meaning of type 1 allows them to be combined to imply meanings of type 2 and express meanings of type 3. Any philosophical theory of meaning, if it is to be applied to natural language, must take all of these varieties into account. The fourth common use of the word ‘meaning’ I consider to be incoherent. If one were to ask a layman what, in his opinion, was the most important question for philosophers to answer, he might well reply: ‘What is the meaning of life’? By this, I understand, he does not ask for a dictionary definition of the word ‘life’, nor does he ask what are the intentions of ‘life’. Could he be asking what inferences can be drawn from life? I rather think not. If we were to respond that life means death, as a consequence, or that life means certain chemical substances appeared on Earth at a certain time deep in the pre-history of the planet, as a cause, or even that life means suffering, as an inference from a Buddhist view of human existence, I suspect he would be dissatisfied and ask again: ‘Yes, yes, but what is the meaning of it all?’. A word may have a meaning, an event may have a meaning, an individual may have a meaning, but an abstract concept, what is already the meaning of the word ‘life’, that cannot have a meaning. It is meaningless to suggest that it does. It is possible to argue here that this sense of meaning is synonymous with ‘purpose’ or ‘reason why’. That doesn’t much help, though, because these words are relatives, in this sense, of meaning as intention. For anything to have a purpose, it must have been intended for that end, by someone or something possessing will. The purpose of a bus shelter is to protect waiting passengers from the elements, though it may equally well serve to obscure amorous teenagers from their neighbours. Hikers, or indeed amorous teenagers, caught in the rain may find shelter under an overhanging rock, but that cannot be its purpose, since no-one intended it for that use. I would not like to delve deeply into the philosophy of action, but it is reasonable to assume that human actions, especially those which are carefully considered and consume time and resources, have a purpose, a goal: that is, something is intended by them. Apparently purposeless action is condemned and ridiculed. It is no surprise, therefore, that men seek a purpose to their lives as a whole; and, since intentions can only come from the will of an intender, it is no surprise that they have created beings with will outside of the life that they can observe around them, whose will might be their source. Still, even if one were to accept that the will of God lends meaning to life, there still remains the question of why God wills this and not that, ‘what is the meaning of his meaning this?’ one might ask, probably to be told that, as a mere human, one cannot understand the ways of the divine. The obvious purpose to biological existence which

20

2 The Meaning of Meaning

the world around us makes clear is scorned: human beings are perhaps the only living things for which the natural urges to survive and to procreate are insufficiently noble and an insult to their pride. Ask then for the meaning of life—this is to misunderstand the concept, the word, meaning. We can search for a meaning in life in the sense that if a man is alive it means that his mother bore him and that the grave shall welcome him back into non-life from which he has but taken a brief holiday; but as to a significance lying behind it all which might be grasped and understood like the meaning of a word, a sign, then that is impossible. Life itself is not language. It cannot signify. It seems that the evolutionary development which enabled man to see significance in signs and eventually construct a language, also left him vulnerable to imagining he could see significance in things which were not signs—to look for meaning, for understanding of the world as he had understanding of his thought. Yet what in life is to be understood? What does it mean to understand that life means only death?

2.2 The Indispensable Nature of Ambiguity There will be much to say on the topic of ambiguity in Part III of this work, which deals with fallacies. Before it be consigned mercilessly to the camp of the enemies of clear thought and reason, however, let it be noted that a multiplicity of meanings to every word in a language is essential: it is one of the most fundamental principles upon which the very possibility of language is founded. In short, to borrow an expression from the Programmers: Ambiguity is not a bug, it’s a feature. Any system of signs in which each sign represented a discrete object could not actually be called a language since it simply would not have the requisite flexibility and creativity within it. In a natural language, one sign can cover many objects: ‘dog’ doesn’t mean my dog, it means any and all dogs. If this were not the case, communication would be extremely limited as every noun would be a proper noun and nothing unknown to one of the speakers could be described by the other. There would need to be an immense array of names just to talk about the immediate surroundings: imagine traversing a forest in which every tree had been christened. The same would apply to activities, if my eating and your eating could not be referred to in the same way we should require a colossal number of ‘verbs’. I put the inverted commas around the word ‘verbs’ because I cannot see how verbs, or adjectives, could ever come into being without the phenomenon of ambiguity. The ambiguous nature of words allows us to see and express the similarities between things, whether they be objects, activities or qualities. Ambiguity also allows the creation of the abstract. The abstract concept of justice is whatever it is that those acts which are called just have in common, permitting them to be identified with the same word. This process of seeing similarity and using the same word to refer to it is far more productive than the simple act of naming. It is the process which builds languages, the process which brings forth the wonderful ability to both create and understand original sentences: if I know what a word refers

2.2 The Indispensable Nature of Ambiguity

21

to in one case, I can understand when it is being used to refer to something similar in another case. This same characteristic of language, however, is the root of an enormous amount of philosophical consternation. Aside from simple cases of ambiguous language not being understood because the meaning is unresolved, the creation of abstracts has led to a belief in their existence as something more than words, indeed, this process is so deeply ingrained that the ambiguity is no longer recognised; justice is taken to be a singular concept, living a life of its own, capable of definition, and, importantly, redefinition. We come to consider ambiguous language an error, forgetting that it is ambiguity which makes our language possible at all. The effects of this error are discussed at length in Chapter 9. For now, I shall improve upon what I claimed in the previous section: it is not so much allowing one mark to stand for something else that is mankind’s greatest achievement, it is allowing that mark to stand for many things. When the word ambiguity is employed in the pages that follow, it will generally be in the sense that a word has two distinct uses, sometimes only marginally so, and that a confusion may spring up as a result. The more fundamental multiplicity of meaning as a marker of perceived similarity, allowing names and descriptions to be shared by an infinite variety of objects and occasions should not be forgotten, however, and hangs behind the discussion, a hidden prop which allows it to continue and, from time to time, affects its stance.

2.3 Definitions of Words by Other Words All definitions are, by definition (ironically enough) attempts to force words into outfits at least one size too small, and are accompanied by the discomfort and awkward contortions that process is bound to bring about. That a word needs to be defined in a given context is proof that there is no common agreement on how it is to be used, and that there is a certain ambiguity to it. The definition removes the ambiguity by prioritising one element of the meaning; a process which may be done apologetically by suggesting that one construal is of particular relevance to the work in hand, or dogmatically, by claiming that the meaning being defined is the true meaning and other uses are erroneous or misleading. Much of the practice of philosophy, therefore, involves the deliberate denial of accepted elements of the meaning of words, and the attempt to deal with concepts which do not exactly fit any word by which they might commonly be described. There will be much more discussion of this too in Chapter 9. Definitions are always controversial because they are intended to be so. They are designed to remove one part of the meaning of a word—perhaps the part another man held most dear—so as to suit the party doing the removing: they are always biased, always confrontational. Even if one is defining an entirely new concept, with an entirely new word, for which no-one as yet has a preferred use, one still runs up against doubts as to whether the phenomenon really exists, since no name

22

2 The Meaning of Meaning

was considered necessary for it previously, and if it clearly does exist, whether it is not covered by the meaning of another word, already familiar, and the term is thus redundant. As an example, take the definition of ‘Art’. For a small group of intellectuals it is in doubt, but, my intuition at least, is that for the majority of speakers of English it is not. It has four essential components (maybe more, but not fewer). 1. It is deliberately constructed. (Not a work of nature or chance. An artefact.) 2. It requires great technical skill. (Not just anyone could do it. Talent and training are necessary.) 3. It is in some sense pleasing—awe-inspiring/beautiful/up-lifting. (Subjective and recognised as such.) 4. It carries some deeper form of expression. (There is some reason for its creation beyond its attractiveness.) Those active in the art world, however, have abandoned points 2 and 3, and played rather loose with point 1, over the last century, to reach a notion of art as, more or less, something constructed to express something, unless, of course, it has been constructed to express ‘nothing’. This is not a debate for this volume. Of the famous hanging urinal of Duchamp, the Tate gallery, which is proud to own a replica (!) tells us that: in 2004 Fountain topped a poll of 500 British art experts as the single most influential artwork of the twentieth century, ahead of Picasso’s Les Demoiselles d’Avignon 1907 and Andy Warhol’s Marilyn Diptych 1962. Simple in form but rich in metaphor, the work has generated many interpretations over the years, and continues to be seen as a work that challenges – or, at the least, complicates – conventional definitions of art. (Howarth and Mundy 2015)

This is a well-known, and therefore useful, example of one group attempting to change the meaning of a word to suit themselves. Duchamp’s ‘work’ became famous in the first place because it was refused a place in an exhibition on the grounds that it wasn’t art—well, of course it wasn’t—but Duchamp and his friends were not aiming to create art, they were aiming for ‘art’. ‘Art’ was a new concept, different from art, and, as such, in no way objectionable. The difficulty comes when the general populace is instructed that works which do not fulfil criteria 2 and 3 above are art and that those who construct them are artists, deserving of public support. How does this relate to argument? See the following: P1: P2: P3: Conclusion:

Art is a general good to society.1 General goods should be publicly funded where necessary. Public funding is necessary for the survival of Art. Art should be publicly funded.

Each of these premises is debatable, none is obviously false; the argument is valid. Certainly, every western society does subsidise the arts to some extent with public money, so it is widely accepted by administrators at least. Once it comes to 1 Throughout

this work, arguments will be constructed with premises numbered P1, P2 and so on.

2.3 Definitions of Words by Other Words

23

be applied to specific cases, however, the danger of equivocation appears: is the Art of P1 and P3 the same Art as the Art of the conclusion? If the ‘art establishment’ and the general public have very different ideas about the definition of art, then, even if both groups accept the argument above, there will be grave disagreements about what should actually be funded; and, since the money is provided by the public, the potential for disgruntlement is clear. In giving my own definitions in later chapters, therefore, I acknowledge that each is merely a definition, not the definition. I acknowledge that it does not define a word, a concept, or a practice of use. I accept that my definition is just that: the definition I choose from among a range of options, because it is the definition I wish to work with. My intention is to explain the phenomena which are relevant to that definition, not to your definition of what is apparently the same word. As such, you must take it or leave it, but you may not claim that it is wrong—although definitions can, of course, be mistaken, such as when one defines a table as a species of reptile—that would be to misunderstand the nature of language, as described above.

2.4 Meaning as Used in Argument I have used the phrase ‘stand for’ to express the purpose of one of the ways in which we employ the term ‘mean’. This should not be taken as an acceptance of a referential theory of meaning. Whether a particular word ‘stands for’ the sense, the reference, the idea in my head or something else is not important here, and whatever it is it certainly does not have to be a solid object in the tangible world. It is possible that language grew out of a group of purely referential words, but it isn’t at all obvious that that group itself could have been described as a language. That language must have developed from what was not language is obvious, unless we believe it was created fully formed by a higher power as a gift to humanity. This is an error which should not be allowed to haunt discussion of the beginnings of language, the feeling that the first words were language and that their meaning was the same thing as the meaning of our words today. As verbal structure has become more sophisticated, so has the meaning of that structure, and there is no reason for us to believe that that process has reached an end point: in the millennia to come, our languages may change, will change, and perhaps in ways such as to render them quite different phenomena to the systems with which we communicate today, and with quite a different sense of meaning. This point becomes more obvious when we look at the way in which theories of meaning have moved away from individual words to look at sentences, and from example sentences on blackboards to real utterances with speakers, hearers and a wealth of other contextual information adding to the wider meaning of the language used. Any attempt to study meaning without these elements is ultimately of little use to us. Indeed, it has been claimed that we cannot say that words actually have meaning in isolation: ‘Only propositions have sense; only in the nexus of a proposition does

24

2 The Meaning of Meaning

a name have meaning’ (Wittgenstein 1974: 14, Sect. 3.3). One might say that words are clues to meaning rather than carriers of meaning. Fortunately, a complete philosophical theory of meaning is not necessary for the study of argumentation. To employ an analogy, one does not need to know how an electronic calculator arrives at the correct answers to mathematical questions in order to be able to use one, or to understand the implications of being able to use one. Such ignorance is no obstacle to the identification of incorrect uses either of the tool itself or the answers it provides. In order to achieve one’s purposes, one must know which calculations to perform, and understand the importance of the answers; the process by which the tool reaches those answers is irrelevant, so long as it does. And thus it goes with language. In order to carry out one’s linguistic intents one must be able to use skilfully the tool of language to express certain meanings, and have an idea of how those meanings will be received and considered. The mechanism by which meaning is transferred by words or sentences from oneself to one’s audience is largely irrelevant. Arguing is something one does with language: it is possible because, by and large, meaning works, that is, language works to express meaning. It is vital for the scholar of argument to know what kinds of meaning certain utterances have and how they relate to each other as steps in the expression of a reasoning process; but what it is for a word to refer or a sentence to have sense, these are not questions with which we are bound to engage. Certain writings on the subject are worth a moment’s consideration, however. The quotation above from Ludwig Wittgenstein is taken from the Tractatus LogicoPhilosophicus, representing what is commonly referred to as the early period in his thinking. Wittgenstein is well-known for the turnaround in his ideas between the writing of the Tractatus and the later period Philosophical Investigations, however, in both works he provides numerous insights into the nature of language and is always clear about the importance of language to everything else in philosophy, and, by extension, any kind of reasoning at all. It is in these works that the philosophical basis for much of what is to follow in this volume may be found. A few examples from the earlier work are worth looking at before we move on to discuss the famous conception of ‘language games’ from the Investigations. In the chapter succeeding this more will be said on the implications of utterances and the inferences which may be drawn from particular language items: this all leads from the apparently obvious thought that: ‘A proposition affirms every proposition that follows from it’ (Wittgenstein 1974: 38, Sect. 5.124). The important consideration here is that one proposition may follow from another, and may follow in various ways. A proposition may explicitly express a piece of reasoning, with a complete set of premises, or it may merely imply some or other of them, so that the conclusion may be said to follow from it, but there is a great deal more which follows from everything that is said and the fact that it is said at all. As Wittgenstein goes on to say: ‘if one proposition follows from another, then the latter says more than the former, and the former less than the latter’ (1974: 39, Sect. 5.14), that is, that whatever can be said to follow from a proposition is only part of what follows from it. If we accept the rule that if p then q, then the proposition p can be said to mean p & q, while q means only

2.4 Meaning as Used in Argument

25

itself. The degree to which all deductive reasoning might be considered tautologous, given that the conclusion is contained within the premises, is discussed in Chapter 8. There are also interesting hints at the movement towards the theory of meaning as use. He writes: ‘What signs fail to express, their application shows. What signs slur over, their application says clearly’ (1974: 14, Sect. 3.262), and what is the application of a sign if not its use? He also states that problems with both ambiguity and misunderstood applications of words can be avoided in a logical language by ‘not using the same sign for different symbols and by not using in a superficially similar way signs that have different modes of signification’ (1974: 16, Sect. 3.325). Applied specifically to spoken language the sign here would be the sound made, and the symbol the word meant, as, of course, different words may be represented by the same sound. The key term in the citation is ‘superficially’: if the use were actually the same, there would be no problem: misunderstandings occur when the use appears at first sight to be similar, but, in fact, is not. To summarise what this means in the development of Wittgenstein’s thought, I appeal to a greater authority. David Pears describes the connection thus: In the Tractatus the central point of the picture theory is that the sense of a sentence is derived from the objects named in it and that it is impossible to say what its sense is, adopting some other standpoint from the one from which they were named, because there is nowhere else to go. The sense of a sentence can only be shown in its actual use. It’s a short step from this position to the position adopted in Philosophical Investigations, where continuous practice is substituted for pin-point naming, but the link between language and the world can still only be exhibited in the actual use of language. (Pears 1988: 474)

In the Tractatus, Wittgenstein had written: ‘If a sign is useless, it is meaningless. […] (If everything behaves as if a sign had meaning, then it does have meaning)’ (1974: 16, Sect. 3.328). As Pears notes: ‘It is interesting that the negative aspect of this account of meaning developed so early, while its positive aspect established itself much more slowly’ (1988: 474). Wittgenstein’s starting point in the Philosophical Investigations, however, is not his own earlier attempts at describing language, but the simplistic explanation of language acquisition put forward by Augustine. This theory he describes as: ‘Every word has a meaning. This meaning is correlated with the word. It is the object for which the word stands’ (Wittgenstein 1953: 2, Sect. 1). This conception of meaning is not denied, it is stated merely to be inadequate. What is described would be a system of communication, but not yet a language. When the famous term ‘language game’ is introduced, the first game consists of only such words and while it may serve a purpose, its difference from actual languages is stark. The discussion of what else is necessary for a language and what constitutes its having meaning is both protracted and open to an incredible range of interpretations. In spite of the sections being generally somewhat longer and possessing a greater degree of elaboration than those in the Tractatus, the argument is, if anything, harder to follow. That may be because there isn’t really an argument at all, rather a series of suggestions, questions, and observations. Even sections which really ought to be arguments and are often referred to as such, the discussion of private language, for

26

2 The Meaning of Meaning

instance, cannot be so easily categorised or reconstructed (see Candlish and Wrisley 2014, for an overview of interpretations and responses). There are two important points which I wish to take from Wittgenstein’s thinking in this work, which are of relevance to our overall goal. Firstly, one of the most famous sections runs thus: 43. For a large class of cases – though not for all – in which we employ the word ‘meaning’ it can be defined thus: the meaning of a word is its use in the language. (1953: 20, original emphasis)

It is a terrible mistake to reduce this to the slogan ‘meaning is use’: the text has so much more to say than that. He does not say that the meaning is the use, he says that when we use the word meaning, we are most often referring to the use. That is to imply that there are times when we use the word meaning without this intention, so the two are not identical. This is reaffirmed later, in part two: ‘Let the use of words teach you their meaning’ (220, Sect. 11), where the one cannot be considered synonymous with the other. When we ask for the meaning of a word, phrase or sentence, as was noted at the head of this chapter, we do not always ask for the same thing. Within this work, since arguing is a form of using words, the meaning of words and sentences will generally be taken to be the way in which they are used, but one does not have to construct a theory of meaning which accounts for all uses of the word and all senses in which ‘meaning’ may be understood, in order to do that. The second, equally famous, idea which we shall keep firmly in mind throughout this work, is that, mentioned above, of the language game, a term ‘meant to bring into prominence the fact that the speaking of a language is part of an activity, or of a form of life’ (1953: 31, Sect. 23, original emphasis). Amongst the games listed in that section are ‘Forming and testing a hypothesis’, the essence of arguing. To say that argumentation is a language game is to recognise that it is a use of language which is, in certain ways, distinct from the multivarious other uses of what is apparently the same language to do very different things. Wittgenstein points out that it makes little sense to try to define what a question is, since it is different things in different games, things which merely share a grammatical form, but where form is neither meaning nor function. Questions in argumentation certainly differ from questions in examinations and polite requests in social situations. Indeed, according to David Pears, Wittgenstein saw language games not as ‘precarious systems struggling to meet some external standard of accuracy, but, rather, as self-correcting standardsetters’ (1988: 457–458) which rather brings to mind the concept of argument and the ‘critical discussion’ found in the Pragma-dialectical approach to argumentation, to be discussed in due course. This realisation has the following consequences: in discussing the meaning of utterances taken to be part of an argument, the analyst must be sensitive to the context, the game and all its rules, in which it was originally made. That may well have been an argumentational situation, or again it might not, but not even all situations of explicit arguing are the same: parliamentary debate is played to completely different rules to those for academic debate (see Chapter 5 for further discussion). This is complicated by the fact that, as the chapter subsequent to this will show, all language is to a degree

2.4 Meaning as Used in Argument

27

argumentative. Whether language games can actually take place without any element of argument going on is just one of the interesting questions this study will raise but make no attempt to answer.

2.5 Conclusions of Philosophy This has, fortunately, and of necessity, been only the briefest of skips through the dangerous territory of the philosophy of meaning, causing us perhaps to think too hard about issues only secondary within the current project, but allowing certain uncertainties to be cleared up and dealt with before moving on. The discussion has centred on the work of one philosopher partly because his thought still dominates discussion of these matters, and partly because his conclusions, if they can be called that, are of most use to us in pursuing the goal of understanding language in argument. What, then, are we to make of the conclusions of philosophy as to the nature of linguistic meaning? There are, I believe, two ideas which are to be held in mind through the subsequent discussion as to the nature of meaning: the first refers to the great weakness of language, the second to its strength. The weakness of language is that it is a poor tool to investigate and describe itself. This is, in part at least because words do not tell us what there is, only what we have felt a need to talk about. Any word used to describe the process by which meaning is conferred and transferred will itself be in need of definition, and the meaning of those words used can be no more substantial than what they are generally taken to mean. Take reference as an example. What does it mean to say that a word refers to something? Well it means that it means that. So what is meaning? Well, a word means the thing it refers to. So round we go again. Reference is actually one part of the meaning of meaning. The mistake is to believe that a word like meaning has some deeper, more philosophically satisfying, meaning than the one we give it in everyday use. This is a conceit of philosophers which is repeated and repeated throughout the canon: we all know what it means to know something, and yet the philosopher will demand that there be a definition of knowledge stating precisely what it is in terms of other words. When then one example is found, as in a Gettier case (Gettier 1963), where the word knowledge does not seem to fit the definition, apparently some discovery has been made about the nature of knowledge! All that has been investigated from the beginning is the pattern of use of the word ‘know’, and, since there is no reason to believe that the patterns of use of words will fit exactly with the patterns of use of other words which sometimes seem to be synonymous, there is no reason to be surprised that there are cases where a proposition may be believed, justified and true, yet still not be known. The insistence that language describes what is, rather than functioning as a tool for the expression of the human experience and the convenience of society, is the essential mistake of western philosophy. It is an error which will feature prominently in the discussion of fallacies in philosophy in Chapter 9.

28

2 The Meaning of Meaning

Language does, however, possess a great strength. It does convey meaning. Indeed, one might define ‘meaning’ as being simply that which language conveys. Conveyance implies speakers between or amongst whom that meaning is conveyed. Any assessment of language which ignores its users is simply missing the point (just as assessment of arguments without arguers is unlikely to be illuminating). Whether the beginnings of language, both for humankind and for each individual one of us, are considered miraculous, mystical or our greatest achievement, there can be no doubt that layers of meaning have been added to language as a whole, and are added by the learner each time language is born anew in an infant speaker. Wittgenstein described language as analogous to an ancient city, where the grand suburbs are built around the narrow winding streets of the centre. As languages grow, so meaning grows with them. Every dictionary requires a defining vocabulary. In order to use the dictionary to look up unfamiliar words, one must already understand the words in that vocabulary, although they themselves are also defined therein. A dictionary taken in isolation is one great circular argument: but dictionaries are not taken in isolation, they are used by people who possess knowledge of that defining vocabulary, or at least the great majority of it. That is to say, people who already use those defining words, so can understand descriptions of the use of other words made through them. Those descriptions, however, are always incomplete: they do not really describe, certainly they don’t fully describe, the way in which the unknown word is used. It is easy to spot a word which has been found in this way by a student and inserted into an essay in order to impress, but stands out as strangely out of place or inappropriate. They do, however, describe something, and, by and large, dictionaries work. To say, then, that individual words have no meaning, that meaning is only found in complete propositions is to commit an equivocation. Words have something. That something can be combined to form longer phrases, and understanding of those individual somethings allows us to understand the combined whole, at least in the majority of cases, although the whole may then take on a something of its own. All of this is not the same thing as that which a complete sentence uttered in a particular context by an individual to another individual who is a member of the same linguistic community possesses, yet it is still meaning for all that. Words have meaning: it is precisely all they have. There can be no such thing as a meaningless word: a sign or sound without meaning is not a word—though it may look like one and be called one for convenience. The meaning of a combination of words will, in the normal run of things be a combination of those individual meanings, but that combination, if it be done properly, according to the conventions of the community, will blossom and bloom and develop meaning of a new kind. This meaning is not the same, but it still bears the name of meaning, such is the paucity of language when we need it most. In this work, then, when meaning is discussed, and it certainly will be in terms of ambiguities, equivocations and semantic discordances, it is to be understood on all of these levels, though not always all of them at the same time. The nature of argumentation makes this necessary. For argument has its pragmatic nature as a communicative exercise, and yet, the success or failure of an argument often turns on the semantics of one word used therein. The two sides to the study of argument,

2.5 Conclusions of Philosophy

29

the abstract logic represented by the words and the discursive function they serve, must both be considered, and a concept of meaning which ignores the needs of the one in favour of those of the other will simply not do.

Chapter 3

Language as Argument

In this chapter, I shall endorse the claim that language use, largely and perhaps in its entirety, is argument. Building on the points made in the previous chapters, I shall assert that to say things is, almost always, to mean other things too: and those other things can only be understood through a process of making inferences based on the premises of what one has heard and what one already knows. Speakers know that their words will be taken and used in this way to construct meanings beyond what has actually been uttered, they expect it. This process is a fundamental part of the nature of language and the nature of its use; indeed, the decision to speak at all, as with any act that is not pure instinct, is the result of some internal reasoning. Language use is, for the most part, a deliberate act, undertaken after at least some reflection upon the circumstances, audience, and consequences of one’s utterance. I shall argue that competent speakers are already competent arguers and that any attempt to divide argumentative texts, written or spoken, from ‘ordinary’ language is bound to fail, based as it is on a misapprehension of what language is and what it contains. There are two well-known theories which I shall employ to back these claims: the Gricean theory of implicatures (1975) and Ducrot and Anscombre’s ‘Radical Argumentativism’ (1989).

3.1 Gricean Implicature The starting point of Grice’s theory is that what is ‘implied, suggested or meant’ is not the same as what is actually said, in the sense of the words uttered (Grice 1975: 43). The verb ‘implicate’ is used to cover the three verbs, and any others with similar meaning, that are cited above, and its counterpart noun ‘implicature’ follows from that. This is an important point: an implicature is not something distinct from what is suggested, meant or implied: it is all of those things and more. It has become standard practice to discuss the distinction Grice makes between conventional and conversational implicatures (see Davis 2014, for example), © Springer Nature Switzerland AG 2021 M. Hinton, Evaluating the Language of Argument, Argumentation Library 37, https://doi.org/10.1007/978-3-030-61694-6_3

31

32

3 Language as Argument

although Grice actually refers to conversational implicatures as ‘a certain subclass of non-conventional implicatures’ (1975: 45) rather than putting them in direct opposition to one another. Simply put, conventional implicatures are semantic, they are based upon the decontextualised ‘meaning of the words used’ (1975: 44), while conversational implicatures are pragmatic and can only be understood when the utterance is considered alongside certain contextual circumstances. Grice gives an example of a sentence containing the word ‘therefore’: X is a, therefore, X is b. He then claims that this sentence does not say that X’s being b follows from X’s being a, only that that is implicated, and, hence, that the statement is not ‘STRICTLY SPEAKING, false should the consequence in question fail to hold’ (1975: 45), by which I am far from convinced. The issue may be better illustrated by Davis’s example of an appositive conventional implicature: ‘Speakers who say “Ravel, a Spaniard, wrote music reminiscent of Spain” implicate that Ravel was a Spaniard—they imply, but do not say, that Ravel was a Spaniard. Hence their utterance is misleading but not a lie if they know that Ravel was French’ (Davis 2014: Sect. 2). I should certainly feel that I had been lied to in this situation. Grice may want to make a distinction between what is actually uttered and what is understood, but if the phrase ‘a Spaniard’ here is not a simple ellipsis of the phrase ‘who was a Spaniard’, then it is without any meaning at all in this sentence—a mere random noun-phrase thrown into the middle of the sentence. If we are to take conventional implicatures seriously, we shall be left with no meaning, only implicature. What is important is that sentences cannot be used without their conventional implicatures, while the existence of conversational implicatures will depend on the context of the conversation. Fortunately, this distinction, and indeed the very existence of conventional implicatures is not a problem we need to grapple with. As Davis points out: ‘Conventional implicatures are as much inferences as conversational implicatures’ (2014: Sect. 2). Whether the inference is based on some circumstances surrounding the utterance or merely on our knowledge of the meaning of certain words, it is an inference from premise to conclusion nonetheless. In the passage quoted below Grice seems to go against Davis’s point and suggest that conventional implicatures are accessible by intuition alone—but what ‘intuition’ actually means is no clearer here than in the other reams of writing in which the word is used: ‘The presence of a conversational implicature must be capable of being worked out; for even if it can in fact be intuitively grasped, unless the intuition is replaceable by an argument, the implicature (if present at all) will not count as a CONVERSATIONAL implicature; it will be a CONVENTIONAL implicature’ (Grice 1975: 50). This passage does make clear that, for Grice, conversational implicatures, at least, are the conclusion of arguments, and that they need to be ‘worked out’, even if the case of conventional implicatures is less certain. As Wilson and Sperber point out: ‘An adequate pragmatic theory should also provide some method of recovering the content of the implicatures themselves’ (1986: 242). Given that implicatures are implicated rather than entailed by the original utterance, this process will have to be one of inductive inference, a position supported by authors such as Bach and Harnish (1979), Leech (1983), and Levinson (1983). Wilson and Sperber, however, still see a place for deductive reasoning. They make a distinction between ‘implicated

3.1 Gricean Implicature

33

assumptions’ and the ‘implicated conclusions’ which are deducible from them: the assumptions are only implied, not entailed, by the utterance, but we can use deductive reasoning to move from them to the implicated conclusions. This process of working out presumably involves the making of inferences on the basis of the utterance and other known facts about the speaker, the situation and so on. Every utterance which has a conversational implicature associated with it is, then, one premise in an argument; and if it has many implicatures associated with it, it is a premise of many arguments. If we accept these statements, and I see no reason not to, then the impact for the study of argument is huge—there is no need to go in search of texts containing arguments or seek out certain situations where elaborate argumentation and debate take place—everything we say, or at least a very good deal of it, involves the process of conclusion following by inference from premises. Take a common situation where one person comments to another on the weather. ‘It’s raining’, I say to my wife as I pull back the curtains. It is possible, I would think, that such a statement could be neutral argumentatively, pure observation, so to speak. Yet it seems unlikely—why bother to say something if one doesn’t mean anything by it? More likely, my utterance is a premise leading to one of many possible conclusions: someone should bring in the washing, the picnic will have to be cancelled, we shan’t need to water the garden this evening. In each case the enthymematic premise is a piece of common wisdom: washing shouldn’t be left out in the rain, picnics can’t be held in wet weather, wet gardens don’t need watering, of which we can safely assume the other person is aware. Sometimes these multiple implicatures are unrelated, they may even be contradictory, but in other cases a whole range of implicatures is necessary for the interaction to be felicitous. If I write: I shall now define argumentation, I am implicating the conclusions that it is possible to define abstract concepts, that it is possible to define argumentation, that it is necessary to define argumentation, and that I am capable of making such a definition. This list is not exhaustive, but it is enough. Each of these claims is based on unstated premises as well as what I have actually written; each is susceptible of counter-argument. The first relies on assumptions about language, the second on assumptions about argumentation, the third on assumptions about the current state of the field, and the last is an auto-appeal to authority. What is interesting in this case, as opposed to cases where there are various possible, but perhaps contradictory or at least unconnected implicatures to choose from, is that if any of those assumptions is questioned, my next sentence is likely to be disregarded.

3.2 Conversational Maxims Since the same utterance can clearly give rise to a number of implied standpoints, some check on what implicatures can reasonably be made, needs to exist. If I am to be held to support every possible conclusion to which my statement might be considered a premise, I shall be reluctant to speak at all. Indeed, the unfair assertion of beliefs in others on the basis of implicatures they did not intend to be made is a common

34

3 Language as Argument

occurrence and a favoured tactic of those who would have others condemned by their own words.1 As so: Detective: You broke into the bank and shot the guards. Suspect: I didn’t shoot anyone! Detective: Aha! So, you admit to breaking into the bank. For Grice, these implicatures are conversational because they are subject to his maxims of conversation. The interpretation we should put, we can fairly put, upon the words uttered by another is that which best fits with his abiding by the conversational maxims: which are of Quantity, Quality, Relation, and Manner. That means that we assume that the speaker is giving us all the necessary information and no more, that the information is correct, that it is relevant to the conversation, and that it has been delivered clearly (Grice 1975: 47). This division is considered unnecessary by Wilson and Sperber, who replace the four categories with a simple test of Relevance: ‘all of Grice’s maxims can be replaced by a single principle of relevance—the speaker tries to be as relevant as possible in the circumstances’ (Wilson and Sperber 1986: 247). In the example given above, it is reasonable to think that the suspect felt the charge of murder was the one which it was more relevant to deny instantly, and it is, therefore, unfair to take his lack of response to the lesser accusation as an admission of guilt. Taking the principle of relevance, we can go deeper. Even if my utterance is purely expressive, and I don’t intend anything more to be understood by my audience, my choosing to make that utterance is still the outcome of a reasoning process about what it is relevant to say at this point in time to this person or these persons present. My utterance of ‘It’s raining’ implies the argument: P1: The state of the weather is relevant to the audience. P2: My statement ‘It’s raining’ refers to the state of the weather. Conclusion: My statement is relevant to the audience. And the more general principle: P1: My statement is relevant to the audience. P2: Relevant statements, all things being equal, are worth making. Conclusion: My statement is worth making. Which also works only on the assumption that we accept that what is worth saying should, if no other reasons intervene—a vow of silence, a sleeping baby—actually be said. Learning to use and understand language, then, is learning how to argue. In order to understand what other people are really saying, we have to be able to follow their implicatures, which allow us to follow their reasoning. In order to assess what they are saying, we have to be able to judge whether that reasoning is good or not, whether those implicatures are really justified. Most importantly of all, in order to speak 1I

have described a series of fallacies identified by Ch. Dickens which rely on this mechanism, in Hinton (2020a), and discuss them briefly in Chapter 7, below.

3.2 Conversational Maxims

35

competently within our language community we need to make sensible inferences and be aware of which premises we need to openly express and which can be taken for granted by our interlocutors. We should also be aware of what conclusions, what implicatures, might be reached on the basis of what we have said, especially those which we do not wish to support. Such realisations are often followed by explanations and a little back-tracking: Guest: I can’t wait to get home. Host: Oh really? Guest: Not that I haven’t had a great time here, and, of course, you’ve been wonderful, and it’s such a lovely house…. When we are not aware of these unwanted implicatures we are likely to come across as rude, brash, or at least, conversationally clumsy, and the English expression ‘when you’re in a hole, stop digging’ may be employed to bring an end to our embarrassment.

3.3 Implicature in Argumentation Research The importance of Grice’s implicature theory has been considered by argumentation theorists, but perhaps not as much as it might have been. Snoeck Henkemans (2014) provides something of an overview of the influence of speech act theory on argumentation study, with some discussion of the role of implicatures. Fabrizio Macagno has also written on the subject, alone (2012) and with Douglas Walton (2014), of which more below. Macagno makes the reasonable claim that ‘implicatures represent forms of inference aimed at retrieving the speaker’s intended meaning and reconstructing his or her dialogical plan’ (2012: 239) and goes on to set out argument schemes for these inferences. He follows Green (2010) in characterising implicatures as inferences to the best explanation, which fits well with the test of relevance described above—the best explanation of the speakers intentions is the one which brings the most relevance to what has been said. Macagno focuses largely on cases where that relevance is not immediately obvious, and a certain statement may seem out of place: ‘conversational, or rather dialogical, implicatures involve a conflict of presumptions that trigger a two-step process of reasoning from best explanation aimed at finding an explanation for the apparent communicative failure’ (2012: 259). Another way of looking at relevance is to talk, more colloquially, of making sense. In order for what someone has said to me to make any sense, certain other things must be true; at least, certain other things must be believed by the speaker. When a man says: ‘One of my daughters is a doctor’, he must have more than one daughter for his statement to make sense. When I ask someone to come for a walk and she replies: ‘No, thanks, it looks like rain’, I am to accept that she must believe that walking in the rain is impossible or unpleasant or simply not for her, although I know that it may actually be my company she objects to. If she had replied: ‘No thanks, all cats are

36

3 Language as Argument

mammals’, I wouldn’t have been able to reach any explanation, any set of implied beliefs, which would give sense to her utterance. In the kinds of examples that Macagno concentrates on, the implicature needs to be worked out more carefully than usual and may even run counter to the prima facie reading of the sentence. He describes the situation where the captain of a ship writes in the log that the first mate was drunk, and the mate responds by writing in the log that the captain was sober. Since the ship’s log is used to record events of note, it only makes sense for the first mate to write there that the captain was sober if the captain’s being sober is a remarkable event: a remark which at first sight reflects well on the captain actually only makes sense if it is an implied criticism. The overall implication of the two log entries is that the mate was drunk on one occasion, and the captain was sober on one occasion. As Macagno points out, we can further understand from this that: ‘The move of the first mate should be read not as an attempt to report a fact, but as an effort aimed at attacking the captain’ (2012: 240). That is to say, there is an implicature to the implicature of the statement. While such deliberately arch uses of implicature certainly exist, they are very much special cases. Macagno is interested, however, because his ‘argumentative model wants to show the implicit mechanism of reconstruction of the purpose of a dialogue move in new or not prototypical situations’ (2012: 262). The reconstruction of the purpose of a speaker’s statement is a second step, coming after the reconstruction of any linguistic or factual implicatures. Understanding that purpose allows us to see what move in any ongoing argument dialogue is being made. If we understand only that the captain was sober, we think well of the captain; if we understand the implicature that his being sober was an unusual event, we think badly of the captain; but if we understand that the mate was deliberately trying to provoke a negative implicature about the captain with an apparently positive, and factually correct statement, then we know that the mate is not a neutral source in our assessment of the captain and we can ignore, or be cautious of, that negative implicature. Although guessing at the intentions of others might leave us open to error, Macagno suggests that ‘the purpose of a dialogical move can be retrieved on the basis of presumptive reasoning grounded on pragmatic presumptions’ (2012: 262) and there is little doubt that in our normal discourse we do reach conclusions as to what intention lay behind a particular statement on the basis of its linguistic content, the implicatures of that content, and our understanding of the context in which it was made. This is entirely in keeping with the principles of Relevance Theory, since ‘implicatures may be recovered either in the course of processing the explicit content of an utterance or in the course of processing some higher level description of it’ (Wilson and Sperber 1986: 261). In their 2014 paper, Macagno and Walton discuss ‘presumptive implicatures’ and show how they relate to several of the most common argument structures frequently described in the argumentation literature. They state: ‘First, implicatures need to be explained in terms of dialectical relevance. And second, they need to be analyzed as implicit arguments, involving a pattern of reasoning leading from a specific premise to a conclusion’ (2014: 210). They then demonstrate how various conversational implicatures can be reached through the employment of argument schemes such as argument from sign, inference to best explanation, appeal to pity, and analogy, while

3.3 Implicature in Argumentation Research

37

the structure of conventional implicatures ‘stems from a different reasoning process, based on a linguistic structure’ (2014: 222). Which fits with what has already been discussed, since if conversational implicatures are to be understood as conclusions to an inferential process, it is to be assumed that that process will mirror those already identified as operating in more explicit forms of argument. Rather than looking for examples of argument schemes at work in the making of implicatures, it might have been more of a challenge to look for schemes which cannot work in this way. Some schemes not considered by Walton and Macagno are illustrated below: Argument from authority A: Fancy a pint? B: John says the pub’s shut this week. This very simple exchange actually masks a fairly long and complex line of reasoning involving the conventional implicature of ‘a pint’, knowledge of the possible establishments for the purchase of pints of beer, and the status of John as an authority on the pub’s opening hours. If he doesn’t want to give up, and assuming it is beer that he wants, ‘A’ may attack either the premise that the pub is the only available outlet: A: We could go to the wine bar down the road. or the authority of John: A: I’m pretty sure it’s next week it’s closing. which would be gentle, or: A John’s a liar. which would be less so. What is undeniable is that while argumentation theorists shape and reshape the form of arguments from authority in their research papers, such examples are being used in conversations the world over continually, and speakers have little trouble in reaching the next move in the dialogue to probe whether the argument, which they have understood through implicature, is a good one or not, without recourse to a list of critical questions. Argument from consequences A: I like the look of him. B: What if your husband finds out? Again, pointing out to someone the possible consequences of an action being considered as a way of discouraging them from performing it is a very common conversational move. There is no need even for ‘B’ to assert what she believes the consequences would be; it is enough, hopefully, to ask a question raising the prospect of those consequences to keep her friend on the straight and narrow. Naturally, implicatures can express the outcome of weak or outright fallacious reasoning just as well. In order for an implied argument to be understood, it does not have to be a good one. Argument from tradition

38

3 Language as Argument

A: Let’s go to Corsica on holiday this year. B: But we always go to Blackpool! The premise that the tradition of visiting the seaside delights of north-west England is in itself a reason for going there again is readily understood, in the sense that we know what ‘B’ is thinking, even if we do not sympathise with the feelings it implies. These patterns of reasoning, which argumentation theorists study and schematise, are essential, built-in parts of everyday language use. Indeed, it isn’t at all clear that they can be separated from language use per se as a particular kind of discourse. There are elements in the analysis of Macagno and Walton that I would quibble with, but their paper is a step in the right direction towards embracing implicatures as instances of argument like any other. What their paper does not address, however, is the wider implication for argumentation theory of the recognition that ‘ordinary’ language, through its endless implicature, and its essentially argumentative nature, is, to a large extent, argument. Conversations involve the making of statements on the basis of what came before, and what one knows from context and experience; and for those moves to be relevant they must be the result of some acceptable process of reasoning on what one’s interlocutor has just said and what one ought best to say in response. As we shall see in the following chapter, where there is reasoning and it is expressed in language, then there is argument.

3.4 Anscombre and Ducrot’s Argumentativity The same conclusions may be reached by following a different path, cleared by the French linguists Jean-Claude Anscombre and Oswald Ducrot. The great majority of their work has been published only in French, however, the development of their ideas is neatly recorded in a 1989 paper translated into English by Anscombre, Argumentativity and Informativity. For the purposes of this chapter, this will be enough to give an overview of their thinking. Indeed, since much of their focus is on detailed analysis of the argumentative value of certain words, it is the general idea rather than the close examination of particulars which is of interest outside the world of French linguistics. They describe their theory as having gone through four stages. The first they name ‘Radical Descriptivism’, in which argumentation and language are considered separately: ‘the linguistic structure of an utterance was not related to its argumentative use’ (Anscombre and Ducrot 1989: 72). This approach, however, was rejected once they had realised that ‘particles like but and even, traditionally analysed as introducing relations between facts, have an argumentative value just like the words generally described as being the very basis of reasoning’, leading to the important conclusion that ‘argumentation, far from being a case by case phenomenon limited to a peculiar form of intellectual activity, is a permanent feature in the use of language’ (1989: 73). Put simply, employing the word ‘but’ rather than ‘and’ to connect two facts is to make an argument about the relationship between those facts: to state that ‘she is blonde,

3.4 Anscombre and Ducrot’s Argumentativity

39

but intelligent’ or ‘he is German, but polite’, is to argue that these combinations are unusual, implying the beliefs that blondes are dim and Germans are rude. The phrases ‘German and polite’ and ‘German but polite’ might be written in the same way in the formal logical notation of an argument, but they are clearly not the same in terms of argumentational value, as one is unremarkable and the other likely to be offensive. This is not only true of logical connectors. One of the key findings which they highlight is the way in which statements with apparently the same informative content can operate differently on an argumentational level. They use the example of ‘as…as’ and ‘the same as’, which Anscombre discussed more fully in his 1982 paper, also in English. Anscombre illustrates how the phrases ‘is as expensive as’ and ‘costs the same as’, which apparently provide the same informational content, cannot be used in the same types of arguments and, indeed, take on rather different meanings when they are negated—‘not as expensive as’ is, as Anscombre points out, generally taken to mean ‘cheaper than’, not simply ‘doesn’t cost the same as’, which could mean that the item in question is more expensive. In this way, we see that there are aspects of the individual lexical items which affect their argumentational quality even if they do not seem to affect the information they purvey. One problem in this description is the assumption that there is informative value to a sentence which exists somehow prior to a full revelation of the semantics of the words contained therein. Is there really any level at which ‘is as expensive as’ and ‘costs the same as’ carry the same information? This, however, is rather the point. What they refer to as the ‘argumentative dynamics’ of the sentence cannot rely simply on the fact which it describes, otherwise we would have to state that the facts described are different in the two cases. Anscombre sums up the position thus: ‘we affirm that the argumentative value of a statement is contained in its semantic value […] there are morphemes whose specific function is to introduce an argumentative component to the semantic value of the statements in which they appear’ (1982: 49). Due to the semantic values of those specific morphemes, any attempt to separate the information contained within an utterance and the argumentative quality of the utterance is doomed to failure. The theory of Radical Descriptivism then develops into what they call Presuppositional Descriptivism. This approach tries to maintain the basic assumptions of Radical Descriptivism: ‘The claim that argumentative sequences are strictly of factual origin remains. But they are no longer based on all the facts referred to by utterances: they only take some of them into consideration. This selection among facts is made on the basis of the linguistic properties of sentences’ (Anscombre and Ducrot 1989: 75). The key change here is that linguistic facts are now part of the argumentative analysis as they have an impact on which facts are taken into consideration ‘namely those which are meant to be asserted within the linguistic structure of the sentence’ (76). The third stage in the development of their thinking they refer to as ‘Argumentation as a Component of Meaning’. Their famous 1983 book L’argumentation dans la langue (Argumentation in language) is based upon this iteration of the theory. They describe the advance made from the earlier stage thus:

40

3 Language as Argument Both stages have in common that the utterance meaning involves factual indications which are determined by a factual component of the sentence meaning. They also both support the claim that the argumentative dynamics lie, at least partially, upon this factual component inherent in sentence meaning. On the other hand, the second and third stage differ in two respects. First, strictly speaking, only the third brings argumentative values into the semantic structure. Moreover, the second only brought out that the semantic structure of the sentence entailed the impossibility of some argumentative structures. The third goes further: it states that at least some sentences (e.g. those with “little”, “a little”, “as… as”) not only can exclude certain argumentative structures, but must also include others. They demand that their utterances must also be used argumentatively and in a given direction. (1989: 79)

This ‘demand’ is brought out by the example of comparing ‘work little’ with ‘work a little’. In factual terms, both phrases seem to indicate an equivalent amount of work—not much, in this context. The realisation which has changed the new thinking from the old is that it is not just the case that we can say ‘Peter has worked a little, he deserves a chocolate’, but not ‘Peter has worked little, he deserves a chocolate’, but rather that since the ‘difference between the two lexical items lies only at the argumentative level. (1) [worked little] must be used to draw the type of conclusion which “Peter has not worked” would lead to, while (2) [worked a little] would argumentatively parallel “Peter has worked”’ (1989: 79). Once the lexical item has been selected, there is no going back, we are forced into a negative argumentative path by ‘little’. Another example of this is discussed by Jayez and Tovena (2008) in their analysis of the French word ‘presque’—‘almost’ in English—and what they describe is equally applicable to the two languages. They point out that: ‘the very idea that presque P is argumentatively aligned with P is misleading’ (227). They show that one cannot use ‘almost P’ as an alternative for ‘P’, or for ‘about P’ by placing it in a sentence with a ‘Few…’ construction: ‘Few car drivers go faster than 120 km/h, about 20%’ ‘Few car drivers go faster than 120 km/h, 20%’ * ‘Few car drivers go faster than 120 km/h, almost 20%’ Jayez and Tovena describe Ducrot’s initial explanation (Ducrot 1972) of the unacceptability of the third sentence as relying on the concept of presupposition. When ‘almost’ is used, there is a presupposition of not-P: ‘¬P is taken for granted. ¬P is all the more likely to be part of the common ground as P is high on some scale’ (2008: 227). They reject this explanation, I think not very persuasively, and the discussion moves on, including reference to Ducrot’s later idea of argumentative scale (Ducrot 1983). What is not in doubt, however, is the principle that there is something in the nature of the word ‘almost’, which is argumentatively different from the word ‘about’ when used to describe quantities, and different from the simple expression of the quantity without either modifier. This difference means that they cannot felicitously be substituted for one another: as in the example of ‘little’ and ‘a little’ above, the presence of ‘almost’ forces us into a certain path as it is not compatible with other directions.

3.4 Anscombre and Ducrot’s Argumentativity

41

Finally, Anscombre and Ducrot hit upon the position of Radical Argumentativism. What is fascinating about this development is that it is hardly a development at all— the basic element has been there all along, implied in what they have said previously. The reason that ‘this car costs the same as that one’ and ‘this car is as expensive as that one’ do not mean the same thing, is that one is a comment on the price, and the other on the expensiveness, or how well the item fits the quality of being expensive. That is to say, they have different frames of reference, and, therefore, necessarily, different argumentative function. This, perhaps, should have been obvious, but it is only now, at the fourth stage, that the ramifications are made plain. Argumentative discourse always involves the application of general principles which we call topoi (pace Aristotle). If we conclude utterance B from utterance A, it is not because A refers to a fact F, and B to a fact G, and that referring to F necessarily brings G in. It is because A brings in F in a way that justifies the application of a topos (or a series of topoi) leading to an utterance B which appears to be a linguistic rendition of G. For us, the meaning of a sentence is the set of topoi whose application is said to be valid when uttered. To choose to utter, in a given situation, one sentence rather than another is to choose to apply certain topoi rather than others in this situation. Here is another formulation: the semantic value of sentences consists in allowing and focussing on facts from an argumentative viewpoint. To choose to describe an object as expensive and not cheap is not to provide some information on its price, but to choose to apply topoi regarding expensiveness rather than cheapness. (Anscombre and Ducrot 1989: 80)

This last sentence seems quite glaringly obvious, but it is the understanding of the importance of this for the study of argumentation which makes the work significant. When they take the idea further they are able to contrast the fourth stage with the third by pointing to two differences: since, previously, argumentative operators were seen as bringing argumentativity into the semantics of the sentence, there could be sentences, without operators, with no argumentative value; and that those with such operators showed two components, the argumentative and the informative. Now, argumentativity ‘is already there in the initial sentences, by means of the topoi which constitute the meaning of the predicates’ (84), and, taken to its logical conclusion: ‘Not only would there be no purely informative sentences, but there would not even be any informative component in the sentence meaning. This does not mean that there are no informative uses of sentences. But it does in fact mean that informative-like uses are derived from a deeper component which is purely argumentative’ (85). One further quotation from Anscombre’s paper rather neatly sums up what this chapter has tried to show, and provides a good indication as to the direction we must take in the later sections of this work: ‘The whole problem of semantic description will then be - in the present case - to calculate the argumentative indications present in the meaning of the statements’ (1982: 52). That those argumentative indications are present, is, I think, beyond doubt; that a semantic description of them is necessary for the study of argumentation I shall attempt to show in what is to follow. As will be the case with most of the chapters of this book, all of this gives rise to more questions than it actually answers. The questions for the theorist, I think, are these: If all language contains arguments, are those arguments somehow different in nature from the more ‘explicit’ arguments usually studied in the field? If so, in

42

3 Language as Argument

what ways and how can we tell the difference? Also, since each utterance we make in support of any particular standpoint will have implicatures beyond the function we are giving it as a premise, is our analysis of arguments not doomed either to lapse into fatal and endless regression, or to simply stop at some arbitrary point? Many of these implicatures might be covered by the agreed common ground, settled, usually implicitly before the argument begins (see van Eemeren and Grootendorst 2004: 60), but what if they are not? I would argue, tentatively, that no, an inference to an implicature and an inference to a conclusion cannot be regarded as different things: implicatures simply are types of conclusions. The question then becomes: is the mere presence of an inference enough to make something an argument? This is where opinions will begin to differ, and much will turn on the definition of argument being employed. The next chapter will consider such definitions and make clear the conception of argument to be employed in the rest of this work, but it can be said at this point that I shall take a very wide view of what constitutes an argument. Looked at in this way, the idea of the argumentative text becomes problematic, although the term may still be used to refer to texts which explicitly set out premise to conclusion relations. An examination of the argumentativity of the language used to express those explicit arguments will have to form part of the assessment of such texts, which is described in the final part of this book.

Part II

Argument

What reason weaves, by passion is undone Alexander Pope, An Essay on Man, Ep. II.

Chapter 4

What Argument Is

In any careful study, it is as well to be sure, and to make clear, what exactly it is that one is studying: in a study such as this, which has much to say about the pitfalls of vague language and the dangers of equivocation when general terms are used with nice distinctions, it is of paramount importance. This chapter aims both to leave no doubt as to how the central concepts of argument and arguing are to be understood in this work, and to show how they have been understood, described and interpreted by other authors. These two tasks are essential to avoid confusion and to illustrate how the appearance of a conflict in thinking may be due only to a difference in conceptualisation. The ‘argument’ of one writer is not necessarily the ‘argument’ of another and it is not the case that one is right and the other wrong; but they must be clear in their definition and they must remain consistent in it.

4.1 Definition of Argument Michael Baumtrog (2018) has provided a much-needed review of definitions and a comparison of how scholars have used certain key words in argumentation theory in differing ways. Below, I follow him in a consideration of some definitions of argumentation itself, as well as in a discussion of the importance and relative neglect of monological or solo argumentation. There is one distinction which he makes, however, which I need to discuss here as it involves two terms which I wish to use differently from the way in which he suggests they be understood. A consideration of his distinction will be useful to avoid confusion and also serve as an example of the way in which definitions can be formed to suit one’s own project. Baumtrog begins by agreeing with a characterisation of conscious reasoning which he finds in Broome (2013) that makes it ‘a phenomenon that occurs in the mind’ (2018: 90). He notes that this reasoning ‘can also be made explicit by putting it into words’ (2018: 90). That suggests two things: that it can, but doesn’t have to be put into words, and, as a consequence, it may be a pre-linguistic stage of thought—it © Springer Nature Switzerland AG 2021 M. Hinton, Evaluating the Language of Argument, Argumentation Library 37, https://doi.org/10.1007/978-3-030-61694-6_4

45

46

4 What Argument Is

isn’t clear whether by ‘put into words’ he means spoken or written, or simply thought in complete sentences. From what he goes on to say, that reasoning ‘can be conducted with multiple participants and is quite often done so’ (2018: 90), one would conclude that reasoning is already in language, but that doesn’t fit with the description of it as ‘a mental activity’ which goes on in the ‘silence of one’s own mind’ (2018: 90). The lack of consideration of when language enters the mental process is perhaps not surprising in a paper on terms in argumentation, and it is not a question which has much exercised argumentation scholars generally; it is, however, of some importance to this work, since I shall argue that many errors in argumentation come about as a result of a failure at this stage. Reasoning, for Baumtrog, is a process of ‘coming to a conclusion’ from some starting point and that may be a ‘joint conclusion’ so that process may be dialogical. That leads him to the obvious question: ‘then what is argumentation?’ (2018: 90). Another definition is required, and the key notion here is that of conflict: ‘A conflict may occur when there is disagreement or doubt about any of the steps in, or the conclusion of, the reasoning’ (2018: 90). It is at this point that argumentation takes place: argumentation is reasoning plus conflict, reasoning accompanied by doubt or disagreement. There are two elements to this with which I am not entirely comfortable. The first is that reasoning and argumentation seem to me to be very, very similar on this account: what is good reasoning will be good argument, and the moment one’s reasoning is questioned it ceases to be reasoning and is transformed into argument. The difference seems to be merely one of attitude not activity. Secondly, if doubt creates argumentation, then I am left wondering in what cases simple reasoning would actually be employed—why reason if the case is not in doubt? One always sets out to consciously consider cases because there is uncertainty surrounding them and at any moment one may cast doubt upon one’s own reasoning process. Also, when doubt or disagreement is thrown upon one’s reasoning by an antagonist, it is not the internal thought that is being questioned, since the antagonist has no awareness of it, but the expression of that thought. It is a frequent occurrence that one may doubt the expression of one’s own thoughts—‘no, that’s not what I wanted to say’ or ‘no, that doesn’t sound right, does it?’ we say, clear in our minds that there is a thought, perhaps a piece of reasoning, which we cannot put into words satisfactorily. Still, I reject Baumtrog’s definitions of reasoning and argumentation not because he is wrong—his explanation is clear and sensible—but because I think they are not useful to me. What he describes as reasoning I do not find to be sufficiently different from what he defines as argumentation that the two need to be separated. Reasoning, as he defines it, might just as well be co-operative or non-adversarial argument: the nuts and bolts of the activities are the same, only the motivation alters. For this reason, I would consider what he calls reasoning to be a type of argumentation, an important type, well-worth individual consideration, but ultimately subsumed within the wider concept; while for him argumentation is a species of reasoning. Baumtrog suggests that for there to be argumentation there must be conflict or doubt, and I rather agree. It seems to me, however, that what he calls reasoning does not take place in cases where there is no doubt. Even where two people work together to ‘reason out’ an answer, and do not compete to persuade each other of the truth, there must have been

4.1 Definition of Argument

47

doubt over what the answer was in the first place, or the process would not have begun. Most importantly, I reject the definitions because they ignore the role of language, which is the focus of my work, and specifically the interplay betwixt language and thought, which I consider an important but neglected part of the study of argument. Suppose I look at your mathematical proof and say ‘I don’t see your reasoning’, or more clearly, ‘I don’t understand your reasoning’. I am not referring to the numbers on the page—I certainly see them, and, if I have a little mathematical knowledge, I understand them as symbols too—what I don’t understand is what they are doing there, why they were so placed, what was the thinking behind using them in this particular way. That thinking does not have to be called reasoning, although I think it generally is, but it certainly can be, and that is what I shall call it. If reasoning is a social process, then not understanding someone’s reasoning becomes merely not understanding their words. The common phrase ‘I don’t understand’ can refer to the expression or to the thought—although one English native speaker would be very unlikely to use it to another in the former case—the two must, therefore, be treated separately. I shall not consider reasoning to be a mental phenomenon. I shall not consider reasoning to be anything at all, as that would be to consider that reasoning exists as something to which the word refers—and always refers. Rather, I shall state that there is a mental process which is pre-linguistic—within the bounds outlined in sections above—and it needs a name. I shall give it the name reasoning as I think that that is what it is commonly called, although I accept that the same term may be used by others to refer to post-linguistic-coding phenomena as well. Those post-coding phenomena I shall call arguments. Still, not all post-coding expressions of thought in language are arguments. Mental processes are opaque and descriptions of them necessarily rely on language items which were either created for other purposes or have meaning only, as it were, at second hand. Much philosophical energy has been expended on the difference between knowing and merely believing, and yet since the workings of one’s mind are no different if the fact one thinks one knows is finally found to be true or false, it can be concluded that belief and knowledge are different words, commonly expressing differing levels of certainty, but, from the point of view of states of mind, only degrees of the same state. Similarly, what the mind is doing when it is reasoning is not a question which can be answered. It is reasoning. We know that it is reasoning because when the thoughts it has come up with are expressed, they are of a certain type, they play certain roles in our language system. In order for a case of mental activity to be reasoning it must involve the taking of some idea, proposition, belief, and in the combination with another idea, proposition, belief, lead to some sort of conclusion. Or at least it must attempt to, since failed reasoning, when we cannot reach an answer, is still reasoning, as much as failing to finish one’s work does not mean one hasn’t been working. We can only have knowledge of some particular reasoning when it is expressed in language, but we may know that it existed previously to that expression, because the expression, if not mere random noise, is the result of that reasoning.

48

4 What Argument Is

There are a number of definitions extant which perfectly well explain how the word argument is used; Tindale has it, reasonably enough, that arguments are ‘characterized by a particular structure, where one or more statements (premises) are given in support of a conclusion’ (2007: 1), and more are discussed below, but, as I have stated above, no definition captures the entirety of any concept. My point is to emphasise that, whatever else they are, arguments are the expression in language of reasoning. Good reasoning plus good coding equals good arguments, but not yet good enough. They must be relevant to serve any purpose, and fit the decorum of the situation in which they are advanced to be considered acceptable and fair. They must, in short, suit the particular process of which they are a part: later, I shall enumerate various processes, or modes of argumentation. I choose, then, to define argumentation, for the purposes of this work, thus: Argumentation is a process involving the expression of reasoning in language. Definitions beget definitions. As I have explained above, reasoning is taken to mean the mental activity which must be coded into symbolic form—language—before it can be expressed. Process is to be understood as the situation, the context, the applicable norms, the whole environment in which that expression takes place. Language will typically mean words, although it could perhaps be substituted by other forms of symbolic expression.

4.2 The Charge of Logocentricism Since the primary concern of this work is with the relationship between words and argument, it will hardly come as a surprise that it is upon argumentation as a linguistic practice that I focus. I discuss only briefly below the interesting and rapidly growing literature on visual and sound arguments, though to the extent to which they rely on symbolic meaning, they can be considered similarly to those expressed linguistically. A few words, however, in defence of my approach from the allegation that I have committed what Michael Gilbert has called the Logocentric fallacy, will be necessary. Gilbert defines the fallacy thus: ‘The Logocentric Fallacy is just the assumption that verbal pronouncements take precedence over other forms and modes of communication, and it is a fallacy because relying on it can often lead us to accept falsehoods rather than truth’ (2002: 26), and he suggests that this fallacy ‘is committed when language, especially in its most logical guise, is seen to be the only form of rational communication’ (2002: 32). His idea stems from the undeniable fact that verbal communication involves a lot more than is usually recorded in a transcription of the conversation. How an utterance is understood may rely on a variety of factors, such as previous knowledge or visual cues, and any attempt to analyse what someone has said without taking into account contextual information of this sort is likely to lead to a misunderstanding of the speaker’s intentions. As much as I agree with what Gilbert suggests, I suspect his argument has fallen foul of the Straw Man fallacy. The following passage makes this clear:

4.2 The Charge of Logocentricism

49

Words, just because they are words, are not inherently clear. The message they contain becomes clear to those who speak the language and are familiar with the issues and contexts. Lay persons reading philosophers often arrive at very odd interpretations just because they are not schooled in the issues and subtleties of the language used. To suppose otherwise is to commit the logocentric fallacy. (Gilbert 2002: 28)

Has anyone ever argued that words, by virtue of being words, are inherently clear? Is there a school of philosophy which maintains that familiarity with the issues and the context are of no import in the understanding of language? Gilbert himself refers for support for his position to the Wittgensteinian theory of meaning discussed in Chapter 2 above, which is hardly a recent or obscure development. Even if we take seriously Gilbert’s warning against relying solely on the surface content of words as an indicator of speaker meaning, I would note that this book, and my entire approach to argument, is not concerned merely with words, but with language taken as a whole. Some of what Gilbert mentions, such as winking when making an utterance, applies only to spoken, not written communication (although a clever writer might include an equivalent to a wink in his text, I suppose). He makes it clear, however, that he is also concerned with written communication: Entire academic industries exist solely because it is frequently impossible to accurately interpret someone’s written words. That is, the mere fact that the words are written down and can be examined and isolated is not enough to ensure precision. (Gilbert 2002: 25)

This strikes me as a somewhat different issue, though, again, I don’t suppose that anyone has ever claimed that the writing down of words does ‘ensure precision’. Written language, which is published with the intention that it be read and studied by those who do not know the author personally, should be susceptible of understanding purely through the meaning of the words alone. Of course, a full understanding of that meaning will require a knowledge of how certain words are used in that context, but that’s an essential part of all linguistic communication. The three examples of disagreement over written texts given by Gilbert are interpretation of poetry, of law, and of philosophy, specifically that of Kant. These are hardly equivalent situations: poetry is a very peculiar case of language use and modern literary theory would suggest that the whole point of the poetic mode is the invitation to personal interpretation, rather than an understanding of what was meant by the writer. When a law is unclear it is because it was either badly written in the first place, or, as is frequently the case, did not foresee the circumstances under which it might be used, making it incomplete. Philosophy which is not clear is simply bad philosophy. When it comes to Kant, it is more the case that he employs a great deal of self-defined technical vocabulary: whether that practice constitutes bad philosophy or not I shall not judge. The point, in all three cases, however, is not that words cannot properly convey the full meaning of the writer, but that they may not do so because of some quality of the writing: the poet deliberately uses words in an unexpected way, the legislator is needlessly vague, the philosopher equivocates. In none of these cases would some extra circumstantial knowledge allow a better interpretation of the text. If the act of putting language down on paper really did ensure that the meaning was clear, there

50

4 What Argument Is

would be few examples of linguistic fallacies in written texts: that is not a claim I have seen made by anyone, scholar or layman. I do not think, therefore, that Gilbert’s charge of logocentricism applies to my definition of argumentation, since I take language to mean not just the words used, but to include the other factors which are always present in the taking of meaning from an utterance, and without which such meaning would not exist. I suspect that in moments of clear and careful thought, the charge does not apply to anyone, however, it may be true that we are all occasionally in danger of committing the logocentric fallacy in particular cases where we fail to realise the possible importance of extralinguistic information and simply take words as constituting the entirety of a message. If nothing else, Gilbert’s paper serves as a useful reminder that language and the process of linguistic communication, consist of far more than merely words and their exchange. I should like to make two small points and one larger one about the inclusion of visual and other types of argument within the field of informal logic studies. The first, which will be controversial, is that it strikes me that a good deal of what is called ‘visual argument’ is only verbal argument which has been visualised. As a result of that, while an image can be understood, its meaning is only clarified and a response given through reinterpreting the image linguistically. Certainly, there are people who claim to think in images (see Grandin 2006) but can they really think of arguments in this way? Are they reasoning? Isn’t it the case that every political cartoon, for instance, begins life as a proposition expressed in language, which is then turned into a picture? The second small point is that pictures can express more sophisticated content through the use of well-known symbols, that is, signs with meaning. If this is the essence of visual argument then it sits happily within my theory: visual arguments are reasoning expressed through the language of signs, and that codification into signs is no different, or at least not different in any way which disturbs its being treated as the same process in argumentation theory, from the codification into words. The two examples given in one of the earliest papers on visual argument, by Birdsell and Groarke (1996) illustrate these points. The first shows a picture of a fish smoking a cigarette and being caught on an angler’s hook. The image makes use of a caption ‘don’t you get hooked!’ (1996: 2). Firstly, the pun only works because of the use of the phrase ‘get hooked’ to mean ‘become addicted’. The visual argument is, therefore, reliant on language for its purpose, even if the caption is removed. Secondly, and following from this, it is clear that the real argument is one which has been made linguistically and then illustrated through the picture. There is no force to the conclusion—don’t get addicted to cigarettes—unless the image is reconstructed in words. The second example is a better one and does not need any caption. It shows an overweight artist painting a scene in which an innocent reality is portrayed as sinister brutality. The point being made is that the capitalist controlled media and complicit politicians of the West misrepresent the benign reality of the Soviet Union. The cartoon is clever and forceful, but it is only meaningful if certain signs within it are interpreted correctly: the hammer and sickle is replaced with a skull and crossbones, a workman’s hammer with a knife, and the obesity of the artist marks him as

4.2 The Charge of Logocentricism

51

a capitalist. The picture, then, employs a system of symbols just as surely as any text does. If this is the meaning of ‘visual argument’ then it presents no difficulty to my approach. The larger point which should be noted is concerned with what the concept of a multi-modal argument says about the concept of argument itself. There is no doubt that arguments can be illustrated, supported, or even rebuffed, with pictures, sounds, stories, and many other artefacts. The question is, should they be considered a part of the argument itself? In order to answer that question we must first decide upon what an argument is precisely. Once that has been established, we may look at how arguments relate to the activity of argumentation.

4.3 Argument and Argumentation I have already defined both terms, if only implicitly, in my own approach: an argument is the expression of reasoning in language, and when such an argument is delivered within an appropriate context, a certain type of process, then, argumentation occurs. This only goes so far, however. There are two main questions which remain to be resolved: what kind of language must be produced for us to recognise that what is being expressed is reasoning—and that will include considerations as to how much of the reasoning process should, and indeed can, be expressed if that language is to be considered a full argument; and, secondly, what kind of processes qualify as argumentational? A discussion and description of the different modes of argumentation, along with an explanation of what I take a mode to be, is found in the following chapter. Here, I shall be concerned with the general principles that might be applied in order to recognise those modes. There have been many other attempts at defining argumentation. Many researchers have sought to define what it is that they want to discuss, and have chosen to label it argumentation. It should be clear from what has gone before that my definition is not intended in any way to supplant those others—it is neither a rejection nor a correction of anyone else’s understanding of the term. It is, as I have stated, a setting out of what I shall understand the phenomenon I am interested in to be. I believe that it is a definition which will allow me to explain that which I wish to explain, but I do not take it to be a description of some extra-linguistic entity which exists out there in the world independently and has simply been waiting for its correct expression in language. Firstly, let us be clear that the separation of the terms argument and argumentation is essential. Argument is a word with a lot of ground to cover: even omitting the popular use of the word to mean a quarrel, it is still asked to stand for both individual premise to inference relationships and the wider exchange and criticism of those relationships. An argument, that is to say, may contain a great many arguments. If we add to this load also the meaning of the process of forming individual arguments and exchanging them within a broader argument, we have three levels of activity, all referred to by the same word. One might properly state that: ‘I lost the argument,

52

4 What Argument Is

despite having the better arguments, because my opponent is skilled in argument.’ This is not a happy state of affairs. This also leads to a difficulty with the idea of metaargument,1 in parallel with that discussed above concerning reasoning, since in every argumentational exchange comments are made and judgements passed on arguments. Should these comments be considered meta-arguments or simply arguments? Metaargument could be used to mean discussion of entire discursive exchanges, or it might mean the sort of statements made in the discussion below, where conceptions of the nature of the process of employing arguments are offered. All three of these activities are regularly engaged in by argument theorists. Given this multiplicity of meanings, referring to the third level, the process of the giving and taking of arguments, as argumentation, seems eminently sensible, although there is more to be said as to what exactly constitutes that process. It would also be useful to cleave apart the other two levels of the term, but far more difficult. Even if we were disciplined enough to use argument only to refer to individual premise/conclusion combinations, and always to refer to the discourse in which these take place as a discussion, argumentative dialogue, or some such construction, there is still the awkward question of where one individual argument ends and another begins.

4.4 The Structure of an Argument Copi, Cohen and McMahon, state that ‘argument refers strictly to any group of propositions of which one is claimed to follow from the others, which are regarded as providing support for the truth of that one. For every possible inference there is a corresponding argument’ (2014: 5). Apart from being rather inelegant, this definition is reasonable enough; yet, it tells us nothing about what sorts of things that group of propositions can and should include if it is to be considered complete. The same can be said for my own attempt at a simple definition which I have used with my students: An argument must contain two things: a set of premises and a conclusion which is drawn, or inferred, from them. Without a conclusion, an argument has no purpose, without premises, it is merely a statement, an unsupported standpoint or opinion. Here too, although the propositions are named as premises, nothing more is said about them. To make this point clearer, it will be well to look at two ways in which arguments can be analysed and written down: the Toulmin system, and Waltonian argument schemes. Stephen Toulmin’s book ‘The Uses of Argument’ ([1958] 2003) is well-known and there is no need here to discuss his ideas in detail. The point of interest to this discussion is his division of premises into various categories. The use of the word premise to cover everything that leads up to the conclusion keeps the terminology simple, but as Toulmin wonders: ‘Simplicity is of course a merit, but may it not in 1 This is discussed at greater length in Section 10.II and in my corpus study into meta-argumentation,

Hinton (2020b).

4.4 The Structure of an Argument

53

this case have been bought too dearly?’ (2003: 89). The first important step is to recognise the difference between what he calls ‘data’ and what he calls ‘warrants’, which to some extent mirrors the traditional division into minor and major premises, but makes it clear what their role actually is. Data are facts which we cite in support of our conclusion and are always stated explicitly, warrants are more general principles which allow us to move from those facts to a conclusion. Very often these warrants are not explicitly stated as they are too obvious: when a friend says ‘you mustn’t come round, I have a terrible cold’, we do not expect him to follow that fact by informing us that contact with cold sufferers should be avoided because of the risk of transfer of the infection. This second premise, which is needed to reach the conclusion that it wouldn’t be wise to visit, is, naturally, another fact; but it is a general fact, not a particular one, a distinction which, though it might obviously be blurred, is certainly relevant. This is something Toulmin readily accepts, urging readers not to ‘be too cut-and-dried in our treatment of the subject, nor to commit ourselves in advance to a rigid terminology’ (2003: 92). One interesting feature of warrants which only comes to light once the distinction is made, is how they relate to types of argument. Any prediction of the future based on a current fact will be reliant on the possibility of employing a warrant connecting past and future events. To move from the data ‘the evening sky is red’ to the conclusion ‘tomorrow will be fine’ we must have a warrant of the form ‘if x is, then y will be’: in this case ‘if the sky is red in the evening, there will be fine weather the next day.’ Anyone who denies the possibility of such warrants denies the possibility of rational prediction. Similarly, any argument which moves from ‘is’ to ‘ought’ requires a moral warrant with a modal verb: John is guilty, the guilty ought to be punished, therefore, John ought to be punished. Again, if we deny the acceptability of moral warrants, we may reach no moral conclusions. This means that anyone who does actually make predictions or advance moral claims is tacitly agreeing that such warrants are justified. Note that the data could be the same in the two cases: the argument could be altered to: John is guilty, the guilty will be punished, therefore, John will be punished. It is the warrant which decides what kind of argument it is, not the data. These thoughts make it clear that the distinction Toulmin draws is an important and a useful one.2 Up to this point, the Toulmin model doesn’t differ much from more traditional approaches, since both data and warrants are the type of premises which would be expected to be included in the setting down of any argument by logicians. He then suggests including two more elements within the full description of the argument, both of which might more usually be excluded, and both of which also lead us towards an illustration of the difficulty involved in distinguishing single argument structures from the broader argument discourse of which they are a part. The first of these elements are qualifiers, since some warrants ‘authorise us to make the step from data to conclusion either tentatively, or else subject to conditions, exceptions, or qualifications—in these cases other modal qualifiers, such as “probably” and “presumably”, are in place’ (2003: 93). Such conditions are either categorised as (Q) qualifiers, or (R) rebuttals. This raises some interesting points: one is that, suitably 2 This

is reflected in the ‘levers’ of the Periodic Table of Arguments, discussed in section 11.III.

54

4 What Argument Is

qualified, all arguments can be made deductive since the uncertainty is brought into the statement of the argument itself. It may not be by deduction that we can move from p to q, but p does make it probable that q is true. Even if q is then found not to be true, the conclusion that it was likely to be true, based on the premises, is not falsified. Toulmin describes the qualifier as an ‘explicit reference to the degree of force which our data confer on our claim in virtue of our warrant’ (2003: 93). The question then is raised: should the type of reasoning, deductive, inductive, abductive, presumptive, be considered an external characteristic which is understood to apply to the argument or should it be written into the argument itself by way of such an ‘explicit reference’? For instance, should abductive arguments always be written with conclusions of the form ‘therefore the best inference is …’? Or should they conclude with clear, positive standpoints which are simply understood to be abductive due to the nature of their premises? This is a point to return to when discussing fallacies in presumptive reasoning. A second point about qualifiers relates to the conditions and exceptions which may be included as rebuttals in the premises. One particularly expects to find such qualification in legal texts, but again, all presumptive reasoning comes with the condition that the conclusion may be altered if more information becomes available. There is also a danger of a never-ending list of possible exceptions of the ‘provided that the world doesn’t end tomorrow’ variety. It might also be argued that with sufficient exceptions practically any argument can become acceptable, if we are prepared to stick to our original position like Reg, leader of the People’s Front of Judea: ‘All right, but apart from the sanitation, the medicine, education, wine, public order, irrigation, roads, a fresh water system, and public health, what have the Romans ever done for us?’ (Monty Python’s Life of Brian). The final element is the backing (B) for the warrant, necessary as ‘there will normally be other assurances, without which the warrants themselves would possess neither authority nor currency’ (2003: 96). At first it might seem that the backing is merely a warrant for the warrant. Toulmin rejects this stating that warrants ‘are hypothetical, bridgelike statements, but the backing for warrants can be expressed in the form of categorical statements of fact’ (2003: 98). These facts are different from the data, like warrants, in that they do not need to be made explicit. There are many questions to be asked about the distinction between warrants and backing, especially in the light of the example Toulmin gives (2003: 102) concerning the religion of a Swede named Petersen: given the data that Petersen is, indeed, a Swede, it can be concluded from the warrant ‘A Swede is almost certainly not a Roman Catholic’, that Petersen is not one. The backing for this statement can be put as ‘The proportion of Roman Catholic Swedes is minute’. As Toulmin notes, expressed linguistically in this way, there is little to tell between the two and they might easily be confused— hence the advantage in setting out arguments as he suggests, where the distinction is clearly preserved. Indeed, it is clear that very often backing is offered in arguments for warrants which are not expressed. Imagine the following everyday sort of dialogue: Susan: I’m sure you’ll fail your exam. Tom: Why?

4.4 The Structure of an Argument

55

Susan: 90% of candidates do. In this case, which I take to be a typical kind of exchange, Susan has omitted the warrant. Her full argument can be written out thus: Data: Warrant: Backing: Conclusion:

Tom is taking the exam. Those taking the exam are likely to fail. 90% of candidates fail the exam. Tom is likely to fail the exam.

This type of arguing is to be expected particularly in cases where the protagonist knows that a challenge to the warrant is likely, and, therefore, skips straight to the backing, forestalling any rebuttal. The reason this is of importance is that a backing fact may well be true, but without the warrant step it does not actually support the inference. There may well be cases where once the backing is established the warrant is accepted without any examination of what that warrant actually is. In the case of Susan’s argument, there may be other factors which are being missed, such as the fact that Tom is well-prepared, is always in the top ten per cent of the class, or has seen the exam questions. Only when the warrant is spelt out clearly, is it possible to tell if the backing does actually support it and if it is strictly relevant to the case. Here, Tom may point out that those taking the exam, like him, who have a very intimate relationship with the person setting the questions are not likely to fail, and the warrant does not apply in his case. For a study into language and argument this point has a vital practical implication. Language constructions, it seems, may disguise the switch from a necessary supporting premise, the warrant, to the factual backing for that premise. In the linguistic analysis of arguments, therefore, it will be necessary to prise apart the two in order to assess whether backing is masquerading as warrant and a step in the inference chain is therefore missing. Toulmin himself discusses how this confusion is present in the logician’s favourite universal premise ‘All A’s are B’s’, a matter I return to in Chapter 9. The overall lesson we can learn from Toulmin’s account at this point is that it is not such a cut and dried matter what counts as part of an individual argument and what elements are better seen as ‘extras’ which may need to be added to it. Whilst everyone agrees that an argument needs premises, Toulmin shows clearly that not all premises are alike. It would be very unfair to suggest that someone’s argument was incomplete because it did not contain every possible exception, or address every rebuttal. Similarly, when putting forward an argument, we cannot be sure which warrants will be accepted at once by our audience, and for which we may have to provide explicit backing. On the other hand, when those types of premises are included they do not appear as external to the argument. The practice of setting down species of argument in argument schemes and accompanying them with Critical Questions does not originate in the work of Douglas Walton, but has perhaps found its fullest expression therein, and certainly the largest number of exemplifications. In his work, critical questions play a very similar role to the ‘extra’ premises included by Toulmin, but here they are placed outwith the

56

4 What Argument Is

structure of the argument itself. Critical questions expose the potential weak points of an argument structure, by which I mean the places at which it may go wrong. Studying such questions might be of use to the student of critical thinking, but it also reveals a good deal about the way in which arguments work for the theorist. It soon becomes apparent, however, that in many cases the critical questions could actually be included in the scheme itself as the types of premises Toulmin listed. For example, the argument from expert opinion, is presented as below with six critical questions (Walton et al. 2008: 310): Major Premise: Source E is an expert in subject domain S containing proposition A. Minor Premise: E asserts that proposition A is true (false). Conclusion: A is true (false). The scheme can easily be rewritten with the questions, or rather the answers to them, now considered part of the argument structure. Major Premise: Source E is a personally reliable, credible expert in subject domain S containing proposition A. Minor Premise: E asserts that proposition A is true (false) based on evidence and consistent with the assertions of other experts. Conclusion: A is true (false). Four of the critical questions have been included, and two, which do nothing more than seek reassurance that the premise, as stated, is true, were already included (field question and assertion question). These concerns might be seen as of a rather general nature, and, for argumentation scholars, somewhat foundational and obvious. They are of importance here, however, because that which is part of the individual argument needs to be expressed in language by the arguer, and that which is not part of it, but is rather implied or understood as belonging to the nature of the process, does not. It may well have been expressed in a general way, by somebody or other, at some point, but it will not be subjected to a detailed linguistic analysis if it is not actually expressed as part of the particular argument under scrutiny. Since linguistic analysis can only be applied to what has actually been said (in a broad sense, including elements discussed in the previous chapter), it may be the case that many analyses will have to be considered incomplete or conducted only tentatively on reconstructed arguments. Certainly, the argument descriptions we have seen in the work of Toulmin and Walton have proved the point that an argumentative discourse is not constructed from a series of premise to conclusion relationships—individual arguments; rather it may involve a refinement and development of the original premise-conclusion pair, meaning that the whole discourse, one wider argument, does actually only feature one individual argument. Fortunately, the distinction of these senses of the word from the sense relating to the activity of arguing can be achieved far more easily through the adoption of the word ‘argumentation’. It only remains, then, to say what precisely

4.4 The Structure of an Argument

57

that is to be taken to mean. My definition has been stated above, but it would be well to get a view of how other theorists, some of whose work I shall refer to in later chapters, have understood and applied the concept.

4.5 Argumentation as Activity Ralph Johnson points out, as we have done above, that argument may refer ‘either to the process or the product of that process’, but it has a ‘proper context: the practice of argumentation’, by which, he says, ‘I understand the sociocultural activity of constructing, presenting, interpreting, criticizing and revising arguments’ (Johnson 2000: 12). He goes on to say that argumentation ‘is an exhibition of rationality’ (2000: 13). This, inevitably, leads to another definition: ‘Rationality is the ability to engage in the practice of giving and receiving reasons’ (2000: 14). So, argumentation is a display of the ability to reason. Which, I think, is unobjectionable, as far it goes. Johnson makes no mention, however, of the medium of that display—language—his book is entitled ‘Manifest Rationality’ and his focus is on the rationality rather than the form in which it becomes manifest. Van Eemeren and Grootendorst begin their definitive description of the pragmadialectical approach with a precise definition of argumentation, which they admit at once ‘differs […] from the way in which the meaning of the word ‘argumentation’ would be described in everyday language’ (2004: 1). They also note that their understanding of the word is deliberately ambiguous between product and process. This allows them both to talk about argumentation as an activity and also to use indefinite articles and make argumentation a countable noun: ‘An argumentation consists of one or more expressions in which a constellation of propositions is expressed’ (2004: 2), something which others might refer to simply as an argument. As I mentioned, their definition is precise: ‘Argumentation is a verbal, social, and rational activity aimed at convincing a reasonable critic of the acceptability of a standpoint by putting forward a constellation of propositions justifying or refuting the proposition expressed in the standpoint’ (2004: 1). To Johnson’s ‘social’ and ‘rational’ they have added ‘verbal’, but they also include a narrow description of the motivation behind argumentation: it cannot be used, as Johnson suggests, ‘to explore the development of a line of thought without committing to the conclusion’ (2000: 14) (that sort of thing, it appears, takes place before argumentation begins), only to support or reject standpoints. In my terminology, pragma-dialectics admits of only one process of argumentation—the critical discussion. Its explanations and conclusions, useful though they certainly are, then, only apply fully to that process, and are incomplete if argumentation is understood more broadly. It is an interesting corollary of the pragma-dialectical insistence that the critical discussion takes place between two protagonists already holding opposing views, that the discussion can be said to be the comparison of two separate processes of reasoning, carried out prior to that discussion, by those taking part. Each participant has already reasoned and concluded, and because their conclusions are in opposition

58

4 What Argument Is

to one another they come together to seek the truth of the matter. One might think of this type of discourse as a kind of meta-reasoning, where one’s arguments, that is, one’s reasoning expressed in language, are subjected to attack by further arguments about that reasoning; that is to say that one’s reasoning is reasoned upon. The possibility of a distinction between reasoning, as some kind of primal inferencing process, and reasoning upon reasoning, a secondary stage of critical consideration of that inferencing, is fascinating, but would surely be impossible to draw in practice. How else can one reach final acceptance of an idea, over which there was enough doubt to make reasoning necessary, other than through the rejection of competing ideas? Yet, there does seem to be some difference between the practice of applying certain, let’s say ‘logical’, principles to basic data, be it physical evidence or our own beliefs, in order to reach an inference, and assessing that practice, in ourselves or others, via a linguistic expression of it, then pronouncing that reasoning good or flawed. Players and referees are all ‘doing’ football, but it is hardly the same activity. Arne Naess defines arguments as ‘appeals to rationality in the light of the facts’ and also grounds his approach in pragmatics and the theory of communication. Argumentation is ‘that element in our expressions which carries the power to convince people in rational discussion’ (1966: 97). More famously, Perelman and Olbrechts-Tyteca characterise argumentation by its motivation and its effects: ‘all argumentation aims at gaining the adherence of minds, and, by this very fact, assumes the existence of an intellectual contact’ (1969: 14) and later: ‘Argumentation is an action which always tends to modify a pre-existing state of affairs’ (1969: 54). Although I agree that argumentation must be viewed as an activity, carried out by arguers, I am less comfortable with the ascription of aims and motivations: I am more interested in what is done, to which I have access, than in why it is done, to which I do not. That, however, is my interest and other authors have other priorities, for which their definitions may well be more suitable. Naturally, I believe my definition has a number of advantages: it permits of three elements to argumentation: the process, the reasoning, and the expression in language. That means that errors in argument, whether accidental or deliberate, can occur in one of three ways: through disregard for due process, through poor thinking, or through imprecise, unclear language. This makes it much easier to formulate a theory of fallacy, since all fallacies will fall into one of the three groups (see Chapter 7). It also allows for certain moves to be considered fallacious in one context but not in some other, since the process of argumentation is a changeable aspect. A number of different argumentational processes, and the ways in which the differences among those processes affect the way in which the arguments expressed within them are assessed, are enumerated in the next chapter, although no strict division or description is adopted. This flexibility of approach is also a useful feature. Some processes are individual, some dialogical, still others multi-logical. Some processes have formal rules—courts of law, public advertisements—others progress simply with the consent of the parties involved. Some processes aim at establishing the truth of the matter, others at reaching a consensus or a plan of action, others at merely persuading an audience to believe or act in a certain way.

Chapter 5

The Modes of Argumentation

Traditionally, there are three fields which study arguments: logic, dialectic, and rhetoric. These were originally distinguished as much by the object of their study as by their varying concerns: logic dealt with reasoning, dialectic with discussion, and rhetoric with the making of speeches. While Aristotle was interested in all three, subsequent centuries saw a parting of the ways with the development of formal logic and its distant alienation from the study of rhetorical figures. One way of characterising the modern field of argumentation study is to see it as an attempt at the reunification of logic, dialectic and rhetoric; of interest in linguistic form and logical structure; of public argument and private consideration. This chapter will be concerned with the different forms an argument may take and a discussion of how those differences affect the business of assessing arguments. Practices of argumentation can be divided in many ways, and each division may tell us something new about arguments themselves. Some divisions may suggest an emphasis on one of the three traditional disciplines in the approach of the author, but none can afford to disregard any of them entirely.

5.1 Wohlrapp’s Three Levels of Argument In his extremely thorough discussion of ‘The Concept of Argument’, Harald Wohlrapp distinguishes what he refers to as three levels of argument; namely, Natural, Scientific and Philosophical. Natural argumentation is what takes place most of the time: ‘It is relatively obvious that, should doubt and dissent occur, we do not simply command and coordinate, but attempt to generate agreement. This is done through the medium of argumentative discourse’ (Wohlrapp 2014: 384). This level of argumentation is very close to ordinary discourse as there are few accepted rules and it is plagued by problems of unequal status, sensitivity to attack, and varying levels of commitment. It is, in short ‘a confusing mess’ (2014: 384). On the positive side, © Springer Nature Switzerland AG 2021 M. Hinton, Evaluating the Language of Argument, Argumentation Library 37, https://doi.org/10.1007/978-3-030-61694-6_5

59

60

5 The Modes of Argumentation

however, Wohlrapp notes that in this practice lies the seed of better argument, the principles of asserting one’s own standpoint and defending it, and of criticising the standpoints of others and calling for supporting evidence. If a little more concern with ‘justifiability and validity’ are added, there is the chance to move up into the second level of Scientific argumentation. The second level ‘is the place for making claims and for validating them with justifications or invalidating them with objections’ (2014: 385). It is, obviously enough, the process by which science is conducted, or, at least, it should be. Wohlrapp notes that the role of argument in science may not be fully appreciated due to a preconception that science is about calculating and investigating, rather than arguing and convincing. Scientific claims, however, are open to validation and justification through their coherence with the agreed framework of that science. When that framework itself comes under threat, personal opinions rise to the surface and scientific argumentation begins to break down; ‘scientific standards dwindle away and we have truly descended to the level of natural argumentation’ (2014: 386). Alternatively, the debate can take an upwards turn and ascend to the highest of the three levels. At the Philosophical level of argumentation there are no bounds upon what may be called into question, even the principles of validity themselves. In order for this to be feasible, however, it must be allowed that not everything can be questioned at the same time: some standards still apply even as others are being doubted. The ideas put forward in philosophical argumentation ‘shake up the sediments of the basis of validity […] The philosopher is wondering about these things we have taken for granted and calls them into question’ (2014: 387). Although this level of debate may seem esoteric to some, and may be the least commonly practised, it has profound effects. It causes the greatest changes in our cultural and social worlds, with consequences which reach far beyond those who do the arguing: ‘The life of humanity as a whole realizes the results of philosophical argumentation’ (2014: 388). Wohlrapp suggests that it is this disparate selection of activities which leads to the diverse approaches to the study of argumentation: it simply isn’t possible to consider all three levels within one theoretical approach. He concludes the relevant section: ‘due to the unique stratification of the subject matter of argumentation, which tends to scale up internally to a theoretical meta-level, argumentation theory actually exists only as the study of argumentation’ (2014: 391). Dividing lines are a difficult business, and, clearly, there would be instances of borderline discourses fitting comfortably into none of these levels, possibly even straddling them all. That does not undermine the case made for recognising three main varieties of argumentation though. What is common to them all is what we call argument; they are all capable of satisfying the various definitions of that phenomenon which were considered above. Yet, they are clearly distinct in their practice and their concerns. I shall also argue later that they can be characterised by the types of fallacy to which they are especially vulnerable. Natural argumentation of the everyday sort is least concerned with fallacy, since it is the least self-aware, and while a feeling of unease may accompany the use of certain argument structures, as to their being unfair, irrelevant or illogical, the lack of clear validation procedures makes it difficult for one participant to appeal to such rules in rejecting another’s claims. For this reason, I

5.1 Wohlrapp’s Three Levels of Argument

61

shall state that the lowest level of argumentation is the particular domain of fallacies of process. That may appear self-contradictory: I have just said that there are no rules of process to be broken. It is, however, the lack of awareness of, and agreement upon, such rules which hampers good debate at this level. Argumentation scholars, from the lofty heights of their third level musings, can still apply the rules of good argument discourse to examples of Natural argument and show where fair procedure has been broken, even when the participants are unable to do so, or simply uninterested. The second or Scientific level, should demonstrate a better level of debate behaviour, where relevant arguments are properly made and can be assessed on the basis of their reasoning. At this level, then, we would expect fewer fallacies of process—though some would inevitably appear—and more fallacies of reasoning. Participants have a good idea of the rules of debate and the rules of validation, but make mistakes with their inferences for various reasons. These may be mistakes of logic or mistakes of inference strength, for example. The Philosophical level of argument corresponds to that highest level of debate at which I contend fallacies of expression are a particular danger. The level of subtlety and sophistication of the arguments employed, as well as the nature of the topics at issue, being questions of the most fundamental kind and often turning upon the precise meaning of a word or phrase, mean that equivocations and other abuses of language are a constant menace. Examples of such are addressed at length in Chapter 9. Wohlrapp’s stratification of argumentation is, then, one which can usefully be applied in the assessment of arguments: an understanding of the level of argumentation being studied informs our expectations of that argumentation and where it is likely to be at risk of error.

5.2 Walton’s Types of Dialogue Douglas Walton (2010) identifies seven types of dialogue which operate as frameworks for argumentative discourse. These can be categorised by three separate criteria: the initial situation, the participant’s goal, and the goal of the dialogue. For example, in the Persuasion type, the initial situation is a conflict of opinions, the participant’s goal is to persuade the other party, and the goal of the dialogue is to resolve or clarify the issue. What all of these elements consist in is established at the opening stage of the dialogue; that stage being followed by the argumentation stage and a closing stage, very much in line with the principles of the Pragmadialectical approach. The other dialogue types listed are Inquiry, Discovery, Negotiation, Information-seeking, Deliberation, and Eristic. Naturally, a dialogue of one type may evolve into a dialogue of another type as it progresses. It is interesting to note that the goal of the individual participating and the goal of the dialogue itself are not the same. So, for example, in Negotiation, the participant’s goal is to get what he wants, while the goal of the dialogue is to find an agreement

62

5 The Modes of Argumentation

acceptable to both parties. Although on one level this seems right, there is something unfortunate about terming the conclusion of the dialogue its ‘goal’. Since only thinking beings can have goals, perhaps a different term would be more appropriate. Not only that, but while it is true that if one enters a negotiation, one wants to get what one wants to get, it is also one’s goal to do so in a way acceptable to the other party: that’s the difference between negotiating to buy a car and just stealing it. To put the point more generally, if one doesn’t have as one’s goal the end point of the dialogue type, why would one engage in it in the first place? The participant’s goal must be to reach that end point. What is referred to here as the participant’s goal, might be better considered the participant’s role. The role of the participants in a negotiation is to push for what they want as part of the final outcome, but their goal is to reach an agreement which will be stable and have a chance of being confirmed and acted upon. The seven types can be divided into two groups: three, Inquiry, Informationseeking, and Discovery, this last having first been added to the original six of Walton and Krabbe (1995) by McBurney and Parsons (2001), involve the search for truth or proof of it. The remainder are forms of conflict: Persuasion, Negotiation, and Eristic dialogue feature conflict between the standpoints of participants, while Deliberation is the discussion and comparison of conflicting choices. These two groups fit the suggestions made in defining argument that it must be accompanied either by conflict, or by doubt. In laying out the characteristics of each type, Walton begins by calling the Persuasion dialogue ‘adversarial’ and suggests that the pragma-dialectical critical discussion is an example of this type. It is contrasted with Inquiry, which is described as ‘cooperative in nature’ (Walton 2010: 14). Since an inquiry is supposed to accept only the firmest of evidential foundations, it may quite satisfactorily end in a conclusion that there is insufficient evidence to decide either way. Inquiries are aimed at finding the truth; this is not the case with Deliberation. Deliberation seeks to establish the best option for action in a collaborative dialogue. Participants are supposed to share all necessary information and their own preferences; they are also required to be prepared to sacrifice their own interests for the common good. This marks a clear contrast with Negotiation, where each participant is trying to get the other to agree to what he considers good for himself, not necessarily for the other side. This can often be achieved by hiding crucial information (the car’s gear box is going to fall to pieces at any moment) or pretending to have different goals from those one actually has (agreeing to miss out on an activity which one wanted to avoid anyway in return for some other favour). Walton is happy to accept the addition of Discovery to his original list on the basis of McBurney and Parsons’s (2001) argument that it differs from Inquiry in that the question which is being considered is set at the beginning of an inquiry dialogue, but only emerges part way through a discovery dialogue. This also applies to the placing of the burden of proof. As Walton puts it: ‘The aim of a discovery dialog is to try to find something, and until that thing is found, it is not known what it is, and hence it cannot be set as something to be proved or disproved at the opening stage’ (2010: 20). I would rather categorise Discovery dialogues of this type as a sub-set

5.2 Walton’s Types of Dialogue

63

of Inquiry than as one of the ‘basic’ argument types, but that is not a matter of great importance. The inclusion of Information-seeking dialogues in the list can also be questioned. Walton justifies it on two grounds. It is not merely an asking and answering of questions to get information, there is also the matter of ‘determining the quality of this information by judging its reliability’ (2010: 22) which suggests the exercise of certain standards of proof. Whether such an exercise qualifies as a process of argumentation is debatable. The second justification is equally shaky: Walton suggests that the requester must ‘provide the responding agent with an argument in order to obtain access to the information requested’ (2010: 22). This argument, however, would presumably be an example of either Persuasion or Negotiation: the fact that an activity may involve the use of argument does not make that activity a type of argumentation itself. The exact make-up of the list, which is never claimed to be exhaustive and does not appear to have been compiled and presented with any ordering in mind, is less important than the observations which come to the fore in its discussion. Although he doesn’t make much of it, Walton employs two key distinctions in his descriptions of the dialogue types. The first, is that some are ‘truth-directed’, and others are not; the second that some are ‘collaborative’ and others are not. Collaborative here should not be misunderstood: all such dialogues require a degree of cooperation between the participants: what is meant is a sharing of interests and removal of conflict. There are, then, on a more theoretical level, four basic types identified here: the truth-seeking and collaborative (Inquiry, Discovery), the truth-seeking but adversarial (?), the non-truth-seeking collaborative (Deliberation), and the non-truthseeking adversarial (Persuasion, Negotiation, Eristic). The question mark in the above list suggests that when we are trying to get what is best for us, rather than what is true, we may either work together for the common good, or against each other for individual gain. However, when we seek the truth, we must cooperate and put aside individual desires and preferences.

5.3 Van Eemeren’s Genres of Communicative Activity While the basis of the pragma-dialectical approach was the theoretical concept of the critical discussion, its further development into Strategic Manoeuvring has seen a greater emphasis put on the types of argument discourse found in real-world communication. In a 2005 paper with Peter Houtlosser, Frans van Eemeren set out three ‘conventionalised types of argument activity’, and claimed that: Unlike theoretical constructs such as critical discussion and other ideal models, which are based on analytic considerations regarding the most pertinent presentation of the constitutive parts of a problem-valid procedure for carrying out a particular kind of discursive task, the various activity types and their associated speech events are cultural artefacts that can be identified on the basis of careful empirical observation of argumentative practice. (2005: 76)

64

5 The Modes of Argumentation

Three of these types are discussed: adjudication, mediation, and negotiation, chosen for their degree of institutionalisation and the transparency of their conventions. Adjudication is a process where a dispute is resolved by a third party to whom the arguments are addressed, it is highly institutionalised and is exemplified by courts of law. Mediation is similar in that it involves a third party, but here that disinterested party acts as a facilitator, helping the two sides to the dispute reach agreement between themselves, rather than imposing a resolution upon them. It is less strongly institutionalised than adjudication, but parties submitting to mediation are expected to follow the procedural rules and finally come to some kind of agreement with their opponent. Negotiation differs from the other two, and indeed most models of argumentation, in that it stems from the conflicting interests of the parties rather than differences of opinion or belief. The format is open, but once agreed upon should be stuck to; rules are not so formally institutionalised, but processes such as haggling over prices have well-known and generally accepted conventions in all walks of life. Although both parties come to a negotiation with the hope of an agreement furthering their own interests, there is no pressure inherent to the discourse activity upon them to make one; they can easily shake hands and return to the initial situation. One interesting point made by van Eemeren and Houtlosser is that argumentation is just one of the tools available to negotiators and it is an interesting question how much argumentation, understood as inferring conclusions from premises, is actually employed in negotiations. In later work (van Eemeren 2010; van Eemeren et al. 2014), there is discussion of ‘genres of communicative activity’ which involve argumentation, specifically those named adjudication, deliberation, disputation, and communion-seeking. These are associated with legal, political, academic, and interpersonal communication respectively. They are also set out in diagram form so as to show how their structure is analogous to the four stages of the critical discussion: each has an initial position, starting points, argumentative means, and a discourse outcome. The initial position category is shared with Walton, but the other three provide a more detailed account than his, and, as I have said, it seems more appropriate to talk of an outcome of the discourse than its ‘goal’. The same four stages are identified in van Eemeren and Houtlosser (2015a), but here the categorisation is by ‘cluster of activity types’ and now adjudication, mediation and negotiation are joined by public debate. The lists of types of activity are clearly stated as not intended to be exhaustive and the slight changes in terminology reflect the ongoing development of the approach. The introduction of public debate is interesting as it is, obviously, public, and therefore more amenable to analysis from the rhetorical perspective, which also has its institutionalised aspects. It is unsurprising given the wider nature of the pragma-dialectics project that the focus throughout these developments is on the constraints and acceptable moves institutionalised in these activities. Starting from this comparative inventory of preconditions for argumentative discourse applying to these four clusters of activity types, it can be shown how these preconditions discipline the conduct of strategic maneuvering in the activity types belonging to a certain cluster. (2015a: 389)

5.3 Van Eemeren’s Genres of Communicative Activity

65

This disciplining of the conduct of types of argument discourse is relevant largely, though not exclusively, to what I have referred to as the process component of argumentation. Naturally, some control over what kinds of reasoning are permissible and how they may be expressed is also exerted, but the perspective for analysis is very much that of acceptability within the rule-framework of the discourse type.

5.4 Dufour’s ‘Rationales’ In his paper ‘Uncontroversial arguments’ (2017a), Dufour raises the suggestion of arguments which do not imply any disagreement or confrontation. This involves drawing a careful distinction between dialogue and dialectic, where the former is understood as essentially cooperative, and potentially uncontroversial, and the latter as inherently antagonistic. Dufour blames the concentration on adversarial behaviour in the study of argumentation at least partly on the ambiguity of the English word ‘argument’. He goes on to list four types of argumentation theory by degree to which they take dialectic to be embedded in the concept of argument, following Finocchiaro (2006, 2013). At one end of the scale are theories which see argument as simply ‘an attempt to support a conclusion with reasons’ and therefore ‘incorporate no dialectical requirement, but do not preclude a dialectical use of argument’; at the other end of the scale lies pragma-dialectics, which contains what he refers to as a ‘hyper dialectical definition of argument’ (Dufour 2017a: 27). Even when an argumentative discourse is seen as co-operative under the pragma-dialectical approach, it is still based around a conflict which needs to be resolved. Dufour accepts that any statement can be analysed dialectically, since we can always invent possible objections, but denies that even argumentative discourse has to be so analysed, noting that a dialogue can just as easily be reconstructed into a monological stream of reasoning. The decision, then, to see all argumentative discourse as adversarial is just that, a theoretical choice, not a necessity. Dufour discusses situations in which there is doubt which leads to an exchange of arguments, even where none of the parties holds any particular opinion, and others where an expert gives reasons to support a standpoint in front of non-experts, who cannot be said to represent a counter opinion. This second type of situation comes closer to the process of explanation and Dufour has considered at length this relationship: ‘argument and explanation are two species of the same genus, which I will call a rationale, understood as the public manifestation of the activity of putting reasons forward’ (2017b: 23). This distinction of the two types of rationale is questioned, however. The paradigmatic situations in which an explanation explains that which we agree about and an argument occurs when there is disagreement, are shown to be insufficient to cover all occasions for argumentative discourse; and, since this is a pragmatic distinction, it suggests that, structurally, arguments and explanations do not differ. Indeed, if one does not know if one’s audience agrees about the existence of some fact or phenomenon, then one does not know if one is arguing or explaining, and one may think one is doing the one,

66

5 The Modes of Argumentation

while actually doing the other. While not denying that the two concepts may usefully be discussed as separate phenomena, Dufour concludes that ‘most of the current approaches to a possible distinction between argument and explanation are on the wrong track or are insufficiently developed’ (2017b: 39). The idea of a class of discourse of giving reasons might even be extended to include other concepts such as justification. It is easy to imagine situations in which one speaker offers an explanation which is really a justification and is actually an argument at the same time. Suppose an employee makes an important investment decision and her boss is doubtful about its benefits to the firm. She is asked to provide an explanation for her actions, in order to justify them, and what she actually does is argue that the decision is in the firm’s best interests. She gives the reasons why she took the decision, she provides the rationale. By asking for an explanation, of a fact which is not in dispute, the boss remains neutral and withholds expression of an opinion that it was a bad or inappropriate decision; the employee, however feels she is being asked to justify herself, and also that the only way to do that is to provide an argument that it was a good decision for the firm, on the understanding that both she and her boss are ultimately driven by that criterion in their actions. It may be that even a successful argument along those lines is not a sufficient justification because she, for instance, had acted beyond her authority or taken unnecessary risks, but that is an interpretation by the boss and has nothing to do with the reasoning she actually expresses. Any attempt to draw distinctions between the three goals, arguing, explaining and justifying is forlorn. Discourse does not, in fact, divide up into neat categories and the use of the words argue, explain, or justify is dictated by the attitude of the speaker rather than anything inherent in the activity, and is certainly not supported by structural differences in what is said. Insistence on such distinctions begins to look like a fallacy of the fetishisation of language, of which much more in Chapters 8 and 9. It does seem as though a good deal of the emphasis on the conflict inherent in an argument is based on purely pragmatic considerations. Dufour puts forward two situations, where there is a discussion of possible reasons and where reasons are offered in explanation, in which there is no conflict, but there is certainly the expression of reasoning within a process; and I think there may be many more. Such situations fit the definition of argument offered in this work, since, although the majority of argumentative processes may be designed to resolve some kind of conflict or difference of opinion, the possibility of processes which do not is not ruled out. The evaluation scheme towards which these chapters are progressing will not, then, focus, much less insist, on the adversarial nature of the arguments under consideration.

5.5 Objections to Functionalism

67

5.5 Objections to Functionalism Jean Goodwin takes issue with the fundamental assumptions which underpin the types of categorisation of argument function of which Walton’s and van Eemeren’s are the best known examples. In her paper, helpfully entitled Argument Has No Function (2007), she makes the case that the way argumentation has been functionalised should be rejected, while not actually suggesting that argument has no purpose: she does allow that: ‘People aren’t insane when they make arguments’ (Goodwin 2007: 70). The function claims to which she objects contain three assertions: that arguments are joint activities (by which she means cooperative), that those activities achieve a social good, and that the norms of argument are those rules which are required for the activity to achieve its function. All three are questioned. A large part of her disquiet stems from what she perceives as a lack of empirical evidence for any of this. It is assumed that argument is a joint activity rather than being shown, and then characteristics of argumentation, not least its normative nature, are based upon this assumption. In his commentary on her work, Patterson describes this as Goodwin’s ‘question-begging charge’ and his counterpoint to it is that: ‘Functionalists, in the main, have not offered an account of what argumentation is. Rather, these accounts are accounts of what argumentation might be, or ought to be, or can be modeled as being’ (Patterson 2011: 4). Goodwin suggests that, in reality, argument is used for a variety of functions which have nothing in common with those listed by theorists, and are not centred around resolution of a point of conflict. She gives examples such as social bonding and experimentation in discourse, but there are clearly many others, such as self-expression (see Sect. 5.6 below), showing off, or simply annoying people. On that last score, she also notes that argument is frequently dysfunctional, promoting confrontation and anger: ‘Hostility is not a social good but a social ill; we do not want to promote hostility by establishing norms’ (Goodwin 2007: 76). Patterson responds by noting that just because argument can be used for different things, it does not mean that it has no ‘primary or root function’ (Patterson 2011: 6). He defends this position by employing the analogy of a man who uses a wrench to beat some metal instead of getting a hammer: the wrench can be used to hammer, but that doesn’t stop its real, or primary, function being to loosen and tighten. In this, I think, Patterson is both right and wrong. He is wrong in the analogy because to choose one of the many uses of argument as the correct use, in the way that the wrench has a correct use, i.e. the one for which it was designed, is to beg the question again—why should that use be primary and not another? Argument has not been designed and manufactured to fulfil a certain defined function, as the wrench has. On the other hand, although the analogy soon breaks down, there is a better way to think of the relationship between the myriad real-world functions of arguments and motivations of arguers, and the ‘primary’ function of argument as a discourse type. It can be argued that argument is capable of playing all those diverse roles precisely because it has a well-understood primary function: because we know argument discourse is intended to resolve doubt or conflict, certain argumentative statements give the appearance of

68

5 The Modes of Argumentation

doubt or conflict and show that we wish to cast ourselves as supporters of one side of a debate, or wish to be considered skilled at such debate. The peripheral functions only work because of our audience’s knowledge of the central one. One other point I shall raise here is a corollary to one I made above about the slightly odd claim that an argument type can have a goal, and that that goal may be different to those of the participants. As Goodwin states, if the rules of the argument are designed to help it achieve its goal, then they may run contrary to those of the participant: ‘Every arguer undoubtedly has his own purposes. But why should he be obligated to so act as to achieve the purpose of the dialogue, especially if he is not even aware of that purpose?’ (Goodwin 2007: 81). There is a good deal more to Goodwin’s argument, particularly in terms of the source of argument norms, and to Patterson’s response, but the discussion over the validity of functionalism has provided me with good reasons to focus on a different categorisation. The argument modes described below refer rather to the ways in which argument can be conducted, not to what the purpose or function of that mode of argument might be, nor to the motivation or goal of the arguer in employing it.

5.6 A Typology of Argument Modes The above sections have provided useful distinctions by which arguments may be categorised: the three levels of Wohlrapp, the qualities of being truth-directed and collaborative from Walton, as well as the stages of argument processes from van Eemeren and his co-authors, and as adversarial or explicatory. All of these are distinctions worth making and I shall refer to them in what is to follow; however, there is a different kind of typology that I wish to add. What I shall call the modes of argumentation combine considerations of discourse types and outcomes with those for procedure and level of argument; there is though something more fundamental to the categorisation and its link to the way in which argument, as pure argument, is employed. I have questioned above whether information seeking or even negotiation can really be said to involve much arguing, it is not at all clear that they actually require, although they may occasionally include, the expression of reasoning. In the descriptions of the modes below, I shall also look to define the role of the meaning of the forms of expression: that is, I shall formulate a typology that allows me to concentrate on the semantic rather than the pragmatic characteristics of the language which is used. The modes of argument described here number eight: Solo, Co-operative, Antagonistic, Formalised, Rhetorical, Eristic, Demonstrative, and Explanatory. As I have made clear before, this is not a categorisation which is intended to be authoritative for all scholars of argument: it is one that strikes me as appropriate and allows me to better formulate my own concerns with the analysis of argument and language.

5.6 A Typology of Argument Modes

69

5.6.1 Solo Argument By this term, I refer to argumentation which involves only one participant, whether or not that participant chooses to speak aloud, write down his thoughts, or keep them out of sight and hearing in an internal monologue. It can certainly be employed at all three of Wohlrapp’s levels. As Baumtrog (2018) explains, solo argumentation is not explicitly ruled out by the best-known theories of argumentation, but it sits in them uncomfortably, and would more easily fit into Dufour’s ‘uncontroversial’ category. Van Eemeren and Grootendorst don’t shy away from the topic stating: ‘The pragmadialectical argumentation theory assumes that, in principle, argumentative language use is always part of an exchange of views between two parties that do not hold the same opinion, even when the exchange of views takes place by way of a monologue’ (2004: 59). However, the talk of disagreement and parties, of antagonists, and of opening stages where ground rules are set out, suggests very clearly the dialogical nature of the activity and the idea that one can create an internal dialogue of that nature stretches definitions somewhat. It is also rather hard to fit solo argument into the activities listed in either Walton’s work or the pragma-dialecticians strategic manoeuvring concerns. Baumtrog also notes that argument is usually dialogical—I have no empirical evidence, and I’m not sure how I would get any, but I disagree. Decision making is not argument, but it can be. There doesn’t seem to be much difference between practical argumentation involving two people or one person. My own mind seems to be a constant generator of reasons, connecting two thoughts to conclude with a third, the majority of which I ultimately reject as other thoughts occur to me. When one enters into dialogical argument without first having thrashed things out monologically, one is likely to embarrass oneself pretty quickly. Against this, Patterson (2011), for instance, argues that argumentation is by definition a social activity. He dismisses the idea that a single thinker can engage in argument with himself and cites Christina Slade’s discussion of thinking skills as essentially social. It is not at all clear, however that Slade’s position is relevant here. She writes: In the longer term the propensity to question and justify becomes internalised, so that an inner dialogue reproduces the moves of reflective discourse. The agonised graduate student, working alone, may well be thinking at high levels. However, the teaching of higher order thinking skills can only be achieved through dialogue. (Slade 1995: 233)

Slade’s interest is mainly pedagogical, and it was established in Chapter 1 that linguistic communication is probably necessary for the development of certain highlevel thinking skills, and doubtless for their improvement and refinement too. None of that shows that, having gained those skills, one cannot employ them without company. Patterson takes it that the description of thinking as an ‘inner dialogue’ shows that the single thinker’s reasoning is just a reflection of social argument, but the phrase ‘inner dialogue’ may equally be understood as metaphorical. I might just as well argue that conversations with others represent a sort of external monologue. Whether it is conceived as reflection of the external, or as a form of argument of its own, the process of solo argument cannot be said to mimic directly any of the external

70

5 The Modes of Argumentation

processes of argument. Solo argument is the least constrained form: anything goes in terms of what can be proposed and considered, but if it is a conscious preparation for other forms it is necessary to consider what the restrictions of those forms are: if one knows one will have to justify one’s decision later on to an audience, one must think of reasons which will be permissible at that stage. Importantly, while there are no real rules of process, the reasoning and its expression in language are subject to the same evaluative constraints as always—poor inferences and vague conclusions are made no more acceptable in argument by not being communicated to anyone else. The role of language in solo argumentation which is kept strictly internal is a difficult topic and takes us back to some of the issues discussed in Chapter 1. It is important to remember that solo argument does not just mean the reasoning part of argumentation. In order to become argument, on my definition, that reasoning must be expressed, thus thinking is only argument if it is done in language. I do not think it controversial to state that when an ordinary language user considers matters deliberatively without uttering a word aloud, he is still perfectly conscious of the thoughts in his head having linguistic structure, although, I concede, it may not be a fully-formed structure. That leads to two important points: firstly that we may commit fallacies of language in our solo reasoning by conflating different meanings of a word and thus equivocating, or employing insufficiently precise notions, for instance, or, more intriguingly, we may reach conclusions which appear to be good, but become obviously less good when we are actually called upon to express them in fully-formed structure. This is a phenomenon everyone will be familiar with and leads to comments such as ‘it sounded much better in my head’, or ‘now I say it, it doesn’t sound right’. The lack of necessity for a fully-formed linguistic expression of reasoning to be communicated to a waiting audience may lead to the process of expression being rushed, and premises, and then a conclusion, being accepted before they have been turned into argumentation proper. Fallacies of expression are likely to be missed, and it is for this reason that even the most intelligent of people are well-advised to talk through their private thinking with a friend before acting upon it.

5.6.2 Co-Operative Argument This is what Baumtrog refers to as reasoning, but separated from purely internal monological argumentation. In co-operative argument there are no opposing standpoints, in the sense of positions which the participants actively champion; rather, the techniques of argument are employed in order to reach one. It may be that the co-operators agree on a joint conclusion, or that their paths diverge at some stage and they begin a more confrontational exchange. Either way, arguments are put forward and critically considered without anyone’s being committed to supporting their conclusions. Participants may consider what kinds of argument would be necessary to convince them of a particular standpoint and discuss what is missing from the

5.6 A Typology of Argument Modes

71

currently available evidence or expressed propositions, thus engaging in a kind of meta-argumentation. Such interactions might also be termed discussion, but they are not quite the same thing as there are certainly many discussions which do not employ forms of argument in order to reach a co-operative solution and where those airing their views do so without any intention of changing them. Discussion is a general word for a type of situation and says nothing about the intentions or practices of those involved. The educational importance of this type of dialogue is, again, stressed by Slade: ‘Reflective discourse in a cooperative group is far from the cut and thrust of aggressive argumentation. The cooperative aspect is essential to the teaching of higher order thinking skills’ (Slade 1995: 227). Van Eemeren’s descriptions of the initial positions of the various argumentative discourse types, even communion-seeking, all focus on some form of conflict or disagreement, rather than an agreed-upon doubt which participants might choose to try to dissolve together. This does not mean that co-operative argument is ruled out, since any situation in which we doubt the truth of A can be seen as a conflict between the standpoints A and not-A, it does, however, lead to the modelling of a non-conflictive dialogue in terms of a more confrontational one. Walton, as we saw, is a little more accommodating of the idea of co-operative argument. It should be remembered, however, that all forms of discourse are to a degree co-operative, since an exchange of views and reasons can only take place with the co-operation of both sides. Even political debates between parties who refuse to treat each other with much respect are founded on the principle that all sides involved agree to co-operation in that particular political system, rather than taking up arms, for instance. This degree of co-operation is not, then, what is being recognised in ‘co-operative’ argument. It makes sense, I believe, to consider this, along with the other modes in this typology, as a more fundamental distinction than those between dialogue types which were listed in the preceding sections. There may be any number of discourse activities which are cooperative and there are considerations which are common to them all. In terms of language use and analysis, one would expect argumentation in this mode to include a good deal of collaborative discussion of terms and their precise meanings. In solo argument one may well be unaware of multiple understandings of a term, in antagonistic argument, described below, one may be reluctant to accept the definition of a fellow participant where it differs from one’s own; when one is cooperating, however, it is part of the process to both reach agreement on the terms used and to consider all the ways in which others might have used and understood those terms. This is useful in both assessing the linguistic evidence being considered and in predicting possible outcomes of one’s deliberations in terms of their influence and effect on other people. Such argument should also be free of what might be called ‘rhetorical tricks’ or deliberate obfuscations.

72

5 The Modes of Argumentation

5.6.3 Antagonistic Argument This might be considered the archetypal form of argumentation, although I am far from certain it is the most common. At least two parties employ arguments to resolve a difference of opinion. Ideally, this resembles the critical discussion of pragmadialectics, but the lack of formally agreed rules will often lead to disagreements about what is and is not fair in this process, potentially causing an argument to descend into a row. It is also questionable as to how often such exchanges actually lead to a change in beliefs of either or any of the parties. The confrontational setting in which arguers ‘take sides’ can be expected to lead to their showing loyalty towards one particular view and being disinclined to accept defeat by giving it up and assuming the position of their antagonist. This is true also of interpretations of language. Each participant is likely to assert that he has understood a particular linguistic expression in the correct way, and that his way of using linguistic expressions is the right one. Arne Naess recognises this problem, noting that: ‘One often hears disparaging remarks directed at those who insist on precise expression’ and that although the need for greater precision in language usually shows itself during an argument ‘all we can hope for in the flush and flurry of debate is some fast repair work to keep the discussion going’ (1966: 62–63). Given that participants in this type of argument are less likely to seek shared understanding of language items, and may even be keen to keep certain parts of their own position as vague as possible, the scope for the committal of fallacies of expression is broad.

5.6.4 Formalised Argument Formalised arguments could, theoretically, be co-operative or antagonistic, although it is likely that the formalisation has taken place as a means of holding antagonistic arguers to certain standards of behaviour which they might otherwise violate. Where the process is more co-operative, as in mediation, the formality is required to maintain that state of cooperativeness and prevent a breakdown into antagonism. In formal argument situations the rules of argument are clearly defined and usually controlled by an impartial moderator. The analytical role of the independent adjudicator is confined to policing the process of the debate and disallowing any fallacies of process. This is the case in courts of law, parliaments and many other forms of debating. Another key difference which this mode has with less standardised forms of antagonistic argument is that while the arguers address each other’s points, there is usually an independent audience which is the real target for persuasion. This might be a judge or jury in a court case, or voters in a political debate. Both politicians and lawyers are employed to put a particular case in an openly biased fashion, they are

5.6 A Typology of Argument Modes

73

not expected to seek truth and are professionally disqualified from acknowledging the superior arguments of their opponent. What differentiates this form of argument from the rhetorical mode, however, is the presence of both the adjudicator and the active opponent. This presence means that participants must take into account, to some extent, the ways in which the meanings of their statements could be understood and exercise a degree of restraint. They may be called upon by the adjudicator or the opposing side to clarify any vagueness or clear up any apparent ambiguity. Whether or not this happens in actual practice depends on the type of situation: in court rooms, a judge or opposing counsel may well ask a witness or lawyer to explain exactly what is meant by a certain phrase; whereas in the British House of Commons, the Speaker’s only engagement with language is that Members do not insult each other using various proscribed words; whether the content of their speeches is meaningful or not is none of his concern, much less whether it might contain a hidden equivocation. Even those making speeches from the opposing benches rarely provide any proper analysis of what has just been said, preferring instead to make their own points, prepared and rehearsed earlier. It is part of the understanding of language in argument of this work that fallacies of language are often the most subtle and difficult to spot upon first hearing an argument. It takes a debater of great skill and experience to be able to respond immediately to any type of fallacy in a speech which he hears for the first time; to notice and highlight a fine equivocation in an argument which is likely to be incomplete and may well have been clothed in bluster and bombast is a rare gift, and to be able to do it without sounding like one is simply ‘splitting hairs’ and thereby losing the support of one’s audience is close to impossible. The presence of the adjudicator notwithstanding, formal argument situations often favour quick-wittedness and rhetorical strategies over well-formed and supported arguments.

5.6.5 Rhetorical Argument Such argumentation is typically one-sided, although, paradoxically, the addressee who cannot usually make a direct reply, is of greater importance than in formalised processes. Classically, rhetorical argument took the form of public speeches, but subtler forms of persuasion have long existed in human interaction, employed by story-tellers, evangelists, and artists of all kinds. In the modern world, advertising is ubiquitous and is certainly the most common form of argument to which we are exposed. In many of these situations, some form of response may be forthcoming: we may vote for a politician, join a religion, change our beliefs, or purchase a product, or refuse to do any of those things; but, as an audience, we do not put forward arguments of our own to support or reject the rhetor. Certain rules do exist, such as advertising standards, which a member of the audience may appeal to if he believes a particular instance is inappropriate, with official agencies playing the role of moderator described above. It is also worth noting that the vastly increased interaction between purveyors of rhetoric and their audience which has taken place over the

74

5 The Modes of Argumentation

last decade thanks to the internet threatens this one-way stream of argument and leaves advertisers and others far more exposed to unsupportive responses to their campaigns. There is little reason to think, at this stage, however, that such responses go beyond a more mob-based process moderator holding companies to account over their adherence or otherwise to fashionable social norms.

5.6.6 Erotetic Argument The intention of this typology was to put forward broad categories of reasoning process, rather than types of discourse, though readers may feel this too subtle a distinction to dwell upon. The modes put forward here are intended to represent the most important ways of arguing rather situations in which arguing occurs, and one of these is what can be referred to as erotetic reasoning,1 that is, reasoning through questions. Questions are, of course, an essential part of dialectic, it is a well-known technique of Socrates in Plato’s dialogues to ask his unfortunate victims to accept that certain propositions follow from what they have already asserted, in order to show them that they are wrong in their assertions, but the employment of questions as reasoning is clearly different from the normal use of assertion and inference from it. There are many occasions when this mode of reasoning appears within situations usually considered to belong to a different mode, and this is often where the rules of the argument allow only questions. Cross-examination in court is an obvious example, where a line of reasoning is pursued through the use of a series of questions designed to reach a certain conclusion; the examining counsel may feel that the answers themselves serve his purpose, or he may them sum them up as an assertion attributable to the witness. Prime Minister’s Questions in the UK House of Commons is rather different since members frequently put forward assertions and criticisms in the grammatical form of a question, when it isn’t really intended as a question at all. It is not their reasoning, to the degree to which they have any, which is erotetic, only their procedure. One particularly interesting type of discourse in which questions are used as arguments is the political interview. Where a celebrity interview might contain genuine questions designed to extract information from the interviewee, possibly information which that person would rather not have revealed, political interviews feature journalists using questions to first establish and then examine the standpoints of those elected to, or running for, office. Since it is an interview, the journalist may only ask questions, not put forward his own argument directly, and it requires considerable skill to make an argument in this way. What is clear from such interviews is that most politicians, even the most successful, are spectacularly bad at dealing with this kind of argumentation. A standard tactic is to simply not answer the question, which can be exposed by a tenacious interviewer who continues to ask it. In a famous example, BBC interviewer Jeremy Paxman asked 1 See,

for instance, Leszczy´nsja-Jasion and Łupkowski (2016), Łupkowski (2018), and Urba´nski et al. (2017).

5.6 A Typology of Argument Modes

75

British government minister Michael Howard the same yes/no question 9 times and received no reply. Paxman then pointed out 3 times that the question had not been answered, to which Howard replied that his long and evasive answer was ‘dealing with the relevant point’ (BBC 1997), which might be an archetype of an avoidance fallacy in question-based debate. The audience will certainly have concluded that the answer was yes, Howard had threatened an official, which had political consequences, but Howard did at least succeed in not making a public admission or denial that might have had legal ones. In his examination of White House press briefings, Alan Partington lists as many as nine strategies employed by spokesmen to avoid answering questions, ranging from open refusal to answer, to entering what he calls ‘rhetorical mode’ which, judging from his examples, seems to mean talking about something other than what the questioner was interested in. The most interesting possibility from the argumentation point of view, and the only one which might be considered an acceptable move in good argument practice, is that of ‘Challenging the question’. Partington’s two examples (2003: 243) are worth examining as they show two ways in which the questioner may commit a fallacy: through asking a loaded question—one which contains within it the implied assertion of a statement, usually controversial—and through asking an unclear question, which the interviewee can legitimately send back for reformulation. Example 1: Q:

The Israeli press is reporting today that Prime Minister Netanyahu has asked President Clinton to defer a decision on Pollard until after the impeachment trial. Is that why we haven’t heard what the decision is, which had been expected last month?

Mr. Lockhart: First off, there are several things wrong with the question. There was never expected to be a decision last month. Go back and look at the facts of the case. Here, Lockhart, the spokesman, explicitly rejects the presupposition behind the question, that an explanation is needed for the late appearance of the decision, by claiming that the decision is not, in fact, late. The formulation of the question, with the assertion that the decision was expected earlier placed helpfully at the end, makes it relatively easy for the spokesman to spot the trap and refuse to answer the loaded question. Example 2: Q:

I guess we’re trying to find out whether the President is trying to keep this under seal, whether his position is to ask the court to keep it quiet.

Mr. McCurry: Sam, there are a lot of problems there. I don’t know what the ‘this’ is that you’re referring to. I don’t know what you think is under seal. This second example has something in common with the first in that the questioner is clearly implying that there is something which the President wants to keep hidden and the spokesman knows what it is. The strategy of McCurry, however, differs from Lockhart’s, in that he offers a more linguistic criticism of the question by rejecting the

76

5 The Modes of Argumentation

use of the pronoun ‘this’, and suggesting that the question cannot be answered in its current form. According to Partington, this kind of tactic is used ‘just occasionally’ (2003: 243). As I suggested above, responding to linguistic errors in one’s opponent’s statements, or questions, in real-time is very difficult.

5.6.7 Demonstrative Argument Another mode which is just beginning to receive attention from argumentation scholars is demonstrative argument. Argument as demonstration does not fit well into any of the other modes: perhaps it might be considered a form of rhetoric, but I shall discuss it separately here in order to make certain points about the analysis of the language involved. By ‘arguing as demonstration’ I mean cases where the act of engaging in argument is itself a demonstration of a fact: of an ability or willingness to argue in that way, for instance. That fact may then itself be employed in further argumentation as a proof or refutation. In their 2016 paper, Jean Goodwin and Beth Innocenti, use examples from the women’s suffrage movement, where the act of engaging in debate and giving reasons for the establishment of women’s voting rights, by women, in itself showed that claims that women were not capable of employing reasoning on serious political issues were unfounded. Crucially, even if their reasoning was rejected and their arguments failed to move their audience, their having made those arguments strengthened their case. Cristina Corredor discusses this phenomenon in terms of speech act theory, describing the advocacy of a certain position as an exercitive speech act: It is worth noticing that for the illocution to be successful it is not required that it effectively exercises influence or gains adherence. Even if the addressees are not moved accordingly, and assuming other conditions hold, it is their recognition of the type of act that has taken place, together with their recognition of the way in which it has affected their mutual normative positions, that makes of the speech act the illocution it is. (Corredor 2019: 26)

A second type of demonstrative argument occurs when reasoning is expressed as a way of asserting aspects of one’s personality or individuality. Again, whether or not one’s argument is accepted is less important than whether or not it is noticed by others and the fact of its being made makes some impression upon them. This practice has received only slight attention (Hazen and Williams 1997; Hample and Irions 2015; Hinton 2016) from scholars of argumentation. It is important to remember that such a mode of arguing exists, however, when we come to assess arguments, and, in particular, the language used in making them. If an argument has been made for reasons other than convincing an audience of its truth or even plausibility, we may still analyse it as a case of good or bad reasoning in the normal way, but should bear in mind that its success or failure does not depend on how strong we find the argument to be, nor even on how strong its actual audience finds it to be, but rather on the strength of the impression it makes. If a Christian argues for a piece of doctrine in order to show the strength of his belief, then it is not the soundness of his arguments that

5.6 A Typology of Argument Modes

77

dictates his success, but the sincerity and passion he exhibits. Emotive and evaluative language which might otherwise be considered inappropriate in a debate can better serve the purpose of the speaker than a calmer, more neutral choice of words. In the case of the suffragettes discussed above, of course, the reverse would be true and allowing their statements to be characterised by emotional language would have allowed their audience to conclude that their prejudices against women were indeed justified. While judging the intentions of others is a dangerous business, some regard for the motivation for putting forward the reasoning is essential in a fair assessment of it.

5.6.8 Explanatory Argument Finally, whilst I do not wish to engage deeply in the debate over the differences between explanation and argument, I have indicated above my sympathy for Michel Dufour’s notion of rationale, the providing of which may be adversarial reasoning or may be explanatory. My own understanding of argument is closer to this broader conception of giving reasons, since I have defined argument as reasoning expressed in language, so I would also suggest as a mode explanatory argument. I would add that explanatory argument does not have to be completely without conflict, since, presumably, it is possible for two people to offer differing explanations. This mode of argument is to be treated the same as the others in the broad pattern of its evaluation, although, of course, certain elements of argument assessment which apply to the relationship between proponents of two sides of the arguments and the moves which they make in their discourse will not apply. All of these modes or processes of argumentation could be used to establish the truth of the matter in a disinterested fashion, though none of them has to be, and several of them, I suggest, very rarely are. Even solo argument concerning beliefs rather than actions (which would be expected to favour the interests of the thinker) may actually seek the explanation or understanding of the truth most suited to the individual’s fundamental worldview and previous commitments rather than look for scientific objectivity. It is a commonplace notion that we are reluctant to admit some things even to ourselves, and can reasonably happily refuse to draw certain inferences from one belief that might affect another, hence the term ‘cognitive dissonance’ to describe the existence of incompatible beliefs within the same mind and the feeling of mild discomfort it provokes.

Chapter 6

Pragmatics, Rhetoric and Semantics

In the introduction to their collection on Argumentation and Language, Oswald et al. cite three categories of research examining relations between linguistics and argumentation: the descriptive, explanatory, and semantic perspectives. Descriptive perspectives are those which ‘aim to pair linguistic formulations with argumentative functions’ (2018: 2), that is they look at how linguistic resources are used to perform certain moves in arguments: these are pragmatic perspectives. Explanatory perspectives seek to say how language can be used to achieve the goals of argumentation, this includes work on links between rhetoric and pragmatics, and rhetoric and argumentation. Semantic perspectives are those where ‘the linguistic system is semantically taken to incorporate an argumentative direction in the sense that units are deemed to carry intrinsic argumentative orientations’ (2018: 2–3). This is the approach of Anscombre and Ducrot, and as we have seen, it is a perspective which looks at the argumentative content of words, rather than the wider role of words in argumentative content. This chapter contains a brief discussion of some of the works in the aforementioned volume, and then describes some of the most important aspects of linguistics-based research into argument. There can be no doubt that in recent years the study of argument as dialectic, heavily influenced by pragmatic theories of discourse, has been in the ascendant. The evaluation procedure to be set out below in Chapter 11 keeps the advances made through this approach in mind, but seeks to give more prominence to issues of semantics than has become the norm.

6.1 Linguistics in the Study of Argument Oswald et al. examine the current state of research under a three-way categorisation: focus on Linguistic markers, Discursive processes, and Cognitive operations. Linguistic markers are of two distinct types: ‘in the first case, the linguistic markers semantically express an argumentative relationship between words, while they help © Springer Nature Switzerland AG 2021 M. Hinton, Evaluating the Language of Argument, Argumentation Library 37, https://doi.org/10.1007/978-3-030-61694-6_6

79

80

6 Pragmatics, Rhetoric and Semantics

shaping an argumentation process in the second case’ (2018: 4). That is, there are the inherently argumentative semantic properties of some words, as described in Chapter 3, and there are words which are used to construct argument discourse, and signal the moves being made. The ways in which language can be used to effect these moves is an important part of research into strategic manoeuvring in argumentation. The study of this first type, the ‘argument-within-language’ approach, is purely semantic and pays no regard to the validity of the reasoning or its relationship to the truth. That makes it an area of linguistic theory which argumentation scholars, particularly those interested in close assessment of the language of argument, should keep in mind, rather than the basis from which to develop a theory of argumentation. Study of the second type of markers is much more closely related to the study of argument, in particular to the study of argument as a variety of discourse, although there is also a line of thought which sees all discourse as argumentative, given the inherently argumentative nature of language. In the consideration of argument as discourse rather than a series of logical inferences to be abstracted from that discourse, a foundational role is played by the Natural Logic of Jean-Blaise Grize. The goal of this is to describe the logic of everyday argumentative discourse in a ‘natural’ way rather than to develop criteria for assessment. A key problem with formal logical approaches is that ‘non seulement qu’il efface les interlocuteurs, mais qu’il vise à les éliminer’ (Grize 1986: 46), that is, they do not simply ignore the interlocutors, they seek to eliminate them entirely. In la logique naturelle, this is not the case: ‘Les raisonnements non formels, en revanche, s’expriment eux à travers des discours au sein desquels destinateur et destinataire restent presents’ (1986: 47) (Informal reasoning, on the other hand, expresses itself through discourses in which the sender and recipient remain present [my translation]). This presence is essential to understanding the argumentative role and force of the discourse. The passage makes obvious the differing conception of an argument within the two perspectives: in one it is a series of statements made by a speaker in a situation, for the other, it is a set of propositions with an inferential relationship and truth values. In order to gauge the deductive validity of an argument, as well as being depersonalised, it needs to be transformed and reconstructed out of its original form, into tidier language, or specialist symbols. For Grize, this move is unjustified, since the presentation of the argument may be crucial to its reasonableness, and certainly to its ability to convince. This stance is essentially dialogical as ‘each argumentative discourse is seen as a proposal made by a speaker to a listener in a specific communicative situation’ (van Eemeren et al. 2014: 483). As the name suggests, natural logic is concerned with the logic of a discourse, since it is concerned with forms of argument, and these forms are schematised in a way which fits the situation and the aim. The success of the arguer depends on this schematisation displaying discourse coherence: the discourse must be receivable, that is, capable of being understood in both style and content; it should be plausible, that is recognisable as grounded in reality; and acceptable, to the audience in terms of the values displayed. These are all necessary for the discourse to be convincing for the listener.

6.1 Linguistics in the Study of Argument

81

Grize sums up the difference between his natural logic and formal logic by noting that in the formal approach there is only one relation, that of implication or entailment, which is determined by the truth values of the propositions; whereas in non-formal reasoning, where natural logic is applied, the basic relation may be of different types: causal, lexical, ideological, and so on, and this explains why conclusions from one context cannot necessarily be transferred to another (Grize 1986: 55). This form of approach is interesting to scholars from all traditions, but Grize illustrates precisely why there is a gulf between the discourse analysis study of argumentation, and the more propositional philosophical method: for philosophers, reasoning is exactly that process by which conclusions can be reached which are capable of transfer to other contexts; even when it is a piece of practical reasoning, applicable only to the present situation, the methodology of reasoning, the standards employed, these derive from the philosophical process of truth-seeking argument. This volume, while acknowledging and respecting the natural logic approach, is concerned with the evaluation of arguments abstracted to a degree from the discourse in which they occur. Elements of that discursive situation are important to the understanding of process, as the context in which argumentation takes place, and elements of it too may be employed in the proper understanding of texts, particularly where reformulation is necessary in order to isolate the argument structure, but the aim is still to regard argument in an epistemological rather than purely communicative role. That some theorists do not accept the legitimacy of such abstraction is noted. Oswald et al. go on to discuss a second category of research, into discursive processes, where argumentation is viewed primarily as a contextually grounded form of activity. They summarise this approach thus: tackling argumentation as a discursive process calls for an analysis of argumentative moves in context, taken at the same time as the local or global verbal context coming before and after the moves, the semiotic context surrounding the moves, the speech context anchoring the moves and the generic context constraining or influencing the production and interpretation of moves. (2018: 10)

There are two ways of looking at research of this type: it could be regarded as discourse analysis, which happens to be of argumentative discourse, or it could be a way for argumentation scholars to gain theoretical insights into argumentation by observing it in practice. There is no doubt that in order to properly understand and characterise the processes of argumentation one should examine discourses taking place within those processes, especially when they are not highly formalised. It might also reveal a lot about the nature of certain fallacies to look at the wider context in which they occur, what type of language signals their presence, and what kinds of moves precede or succeed them. At the same time, there is a clear difference between studying the course which arguments take and studying the strength and acceptability of the reasoning they contain. The language analysis I develop in the later part of this book is an analysis of the language of the argument, rather than of the language surrounding the argument. That is not to deny the importance of that linguistic environment, it is simply a priority dictated by the aims of the study which are focussed on argument as a philosophical concept, rather than argument as an

82

6 Pragmatics, Rhetoric and Semantics

activity, bearing in mind all that has been said about the nature of language in earlier chapters. The third strand of research described by Oswald et al. is that done at the level of cognitive operations. This approach focuses on the activity in the mind of the arguers, the processing of arguments, both by the producer and the audience, rather than the public activity of their discourse. Although this is the arena of cognitive science, the authors explain why linguistics, and specifically the study of the way arguments are formed out of language, is so important to understanding how those processes affect argumentation: If specific formulations are likely to yield specific representations which would vary should the formulation vary as well, and since processing an argument requires the representation of the contents of the premises and the conclusion, then in principle specific formulations are likely to influence the outcome of argumentative processing. (2018: 12)

This, they suggest, is of relevance to argumentation study in two ways: first, understanding construction processes can help in argument reconstruction, where parts of arguments are not made explicit; and, second, a clearer picture of evaluation processes should yield insights into what makes certain arguments persuasive.

6.2 Pragma-Dialectics The pragma-dialectical approach is a way of looking at arguments which is explicitly grounded in linguistic theory. The degree to which that is the case is made clear by looking at the four meta-theoretical principles which are considered to be the necessary basis for allowing both normative and descriptive treatments of argumentative texts. These are ‘functionalizing’, ‘externalizing’, ‘socializing’, and ‘dialectifying’ the objects of argumentational study. Functionalization means that we treat every language activity as a purposive act. Externalization means that we target the public commitments entailed by the performance of certain language activities. Socialization means that we relate these commitments to the interaction that takes place with other people through the language activities in question. Finally, dialectification means that we regard the language activities as part of an attempt to resolve a difference of opinion in accordance with critical norms of reasonableness. (van Eemeren and Grootendorst 2004: 52–53).

All four of these principles categorise argumentation firmly as a species of communication, emphasising the public and social aspects of arguing, and treating it as an activity, not an abstraction. Although norms of reasoning, or at least reasonableness, are mentioned, it is the norms of language use which are most prominent. The principles also place the theory squarely amongst those which assume the constant presence of some Other with whom the difference of opinion has arisen. To reiterate: ‘Argumentation is not just the expression of an individual assessment, but a contribution to a communication process between persons or groups who exchange

6.2 Pragma-Dialectics

83

ideas with one another in order to resolve a difference of opinion’ (van Eemeren and Grootendorst 2004: 55). This approach, it must be said, has been spectacularly successful, and a great many aspects of pragma-dialectics have become generally accepted by scholars in the field. Even though the informal logic movement grew out of a dissatisfaction with formal descriptions of real world argument, it still retained a principal interest in arguments as patterns of inference: this is clearly something very different. The focus in pragma-dialectics is on argumentation as something which people do, part of which only is the exchange of arguments. This is an insight which has been taken on board quite universally and has led to a more rounded understanding of how arguments work and how they may go wrong. Pragma-dialectics has also brought a wider awareness and appreciation amongst those engaged in argument study of the linguistic theories it takes as foundational; in particular the work of John Searle and Paul Grice. Pragma-dialectics looks to integrate the Searlean speech act theory with Grice’s maxims of co-operation, to produce rules of use in accordance with a Principle of Communication. Thus, the Gricean maxim of Quality (Grice 1989), which relates to the truthfulness of what one says, becomes: ‘You must not perform any speech acts that are insincere (or for which you cannot accept responsibility)’ (van Eemeren and Grootendorst 2004: 77). The other rules prohibit speech acts which are incomprehensible, redundant, meaningless, or unconnected to the previous acts in the situation. When the specific factors of the argumentative discourse are factored in and combined with these rules, the result is the list of Rules for a Critical Discussion, which runs to a fully elaborated fifteen in van Eemeren and Grootendorst 2004: chapter 6, and a more accessible list of ‘10 commandments’ described in van Eemeren and Grootendorst 1992a and reproduced in their 2004: chapter 8, which are straightforward prohibitions. A further element is the listing of Searlean speech acts (Searle 1979) as belonging to certain stages of the critical discussion. The four types given are assertives, commissives, directives, and usage declaratives, with a fifth from the original typology, expressives, not being considered to have a role in this variety of discourse. These acts can mostly be used at any stage, but assertives, for instance, feature heavily in the first, Confrontation stage, and not at all in the second, Opening stage. There are also various moves which can be made with each type of act: directives, for example, can be used to challenge an opponent to defend a standpoint, and to request from him an argumentation or a usage declarative (van Eemeren and Grootendorst 2004: 67). This way of looking at argument discourse—as a series of moves rendered by the employment of various speech acts—has clearly influenced work such as that discussed in the previous chapter by Cristina Corredor, as well as leading to the later development of pragma-dialectics into the study of strategic manoeuvring. The pragma-dialectical approach is, quite naturally, concerned with pragmatics and dialectic. That means, in my conception of argumentation, that it is mainly concerned with issues of process and has less to say on reasoning and expression. Because of this, while I think it is an excellent starting point for the assessment of arguments as appropriate or otherwise to the given process, it does not completely fulfil the needs of argument analysis as this work perceives them. Rule 15 is concerned

84

6 Pragmatics, Rhetoric and Semantics

with clarity of expression and states that both parties have the right to request and to perform usage declaratives at any times. It is explained that ‘discussants must formulate optimally and they must also interpret optimally’ which means that ‘a discussant must choose formulations that are comprehensible to other discussants’ and ‘be prepared to replace their formulations and interpretations with better ones’ (van Eemeren and Grootendorst 2004: 156). As one would expect, this refers to the rights and obligations of the parties in their role in the discourse; it does not, however, say anything about the assessment of those formulations or interpretations. It is for this reason that the evaluation process I describe below should not be seen as an alternative to pragma-dialectical analysis, rather as a complementary tool.

6.3 Pragmatics and Inferentialism It is not only the pragma-dialectical approach which has made use of notions from pragmatics in the understanding of reasoning. Jim Mackenzie (2014) traces the roots of inferentialism back to Leibniz and Kant, then forward via Frege and Hamblin to Brandom. Some of the key stages along this journey will be discussed in this section, since, although inferentialism is not frequently referred to in argumentation theory, some version of it seems indispensable to the full and proper consideration of the language of arguments. To begin with, a basic definition: ‘Inferentialism is the view that something’s linguistic meaning is a matter of its inferential role […] A statement’s inferential role depends on what it can be inferred from (its circumstances of application), and what can be inferred from it (the consequences of its application)’ (Mackenzie 2014: 131–132). Given that the crucial relationships in arguments are ones of inference, the possible impact of such a theory of meaning on argumentation study is clear. As a starting point for this way of looking at matters we may consider the rationalism of Leibniz: ‘where empiricists begin with a primitive notion of representation and seek to ground in it whatever inferences are to be recognized [… Leibniz] as a rationalist begins with inference and then explains the notion of representation in terms of it’ (Brandom 1981: 479, original emphasis); and for Kant, ‘the fundamental unit of awareness or cognition is the judgement (assertion)’ (Mackenzie 2014: 134), which makes the understanding a faculty of judgement. When we understand correctly, we have made the correct judgements, we have inferred meaning. This, in turn, gives rise to the notion that words do not have meaning outside of the statements made using them. Frege takes this as fundamental: ‘I start out from judgements and their contents, and not from concepts […] instead of putting a judgement together out of an individual as subject and an already previously formed concept as predicate, we do the opposite and arrive at a concept by splitting up the content of a possible judgement’ (Frege 1881: 16–17). In modern studies of argument, this theme is taken up by Charles Hamblin, with his conception of reasoning dialogue as consisting of ‘commitment stores’ of the

6.3 Pragmatics and Inferentialism

85

participants, the role of which is ‘to provide us with a dialectical definition of statements’ (Hamblin 1970: 265). When we make a statement or assent to somebody else’s, we add that statement to our store until such time as we offer a retraction. The similarity with the speech act based approach of the pragma-dialecticians is noted by Douglas Walton (1993), whose own account of commitment ‘goes beyond the “externalized” notion of commitments as explicit concessions found in Hamblin and van Eemeren and Grootendorst by postulating non-explicit (dark) commitments that have to be inferred by presumption’ (Walton 1993: 94). In all systems, there is a requirement of consistency: clashing commitments derail the dialogue. That may not always be possible in the real world, especially as Walton points out that we may not even be aware of all of our own ‘dark’ commitments. As Hamblin notes: At first sight we would suppose it to be a requirement of the statements in a commitmentstore that they be consistent; but on reflection, we may come to think that, although there does exist an ideal concept of a “rational man” which implies perpetual consistency, the supposition is by no means necessary to the operation of a satisfactory dialectical system. In fact, even where our ideals of rationality are concerned, we frequently settle for much less than this; a man is ‘rational’, in a satisfactory sense, if he is capable of appreciating and remedying inconsistencies when they are pointed out. (1970: 263–264)

That is, we cannot expect total consistency from others or from ourselves, but neither can we defend or accept an inconsistency when it is once exposed. This raises some interesting points for argument assessment. What one is committed to, in this sense, is whatever can be inferred from what one has asserted— even if one does not believe that oneself—as Hamblin points out ‘consistency presupposes the ability to detect even very remote consequences of what is stored’ (1970: 264). These consequences of one’s commitments need to be drawn out in any assessment of one’s position, but that will occur at different levels of analysis. The analysis described in Chapter 11 below would allow for these implications of an arguer’s assertions to be exposed at an initial, first sight of the argument, or at a later stage of reasoning analysis, or, finally, and most likely for those which are ‘dark or ‘remote’, at the stage of deep linguistic analysis. The ability to find these inferential relations, and to find them quickly, also depends on the level of one’s knowledge and familiarity with the topic, and perhaps with one’s skill and experience in dialectic as well: ‘Expertise in a field includes the ability to recognise as immediate implications which those without that expertise need to have explained’ (Mackenzie 2014: 129). I shall close this section with a quotation from Brandom in which he responds to the criticism that since reason-giving is just one function of language, one game which can be played with it, rationalism as a basis for meaning is over-reaching itself. Brandom’s response recalls certain points made earlier in Chapters 1 and 2, and reaffirms the centrality of arguing to the very idea of communication: that our expressions play a suitable role in reasoning is an essential, necessary element of our saying, and their meaning, anything at all. Apart from playing such a role in justification, inference, criticism, and argument, sentences and other locutions would not have the meanings appealed to and played with by all the other games we can play with language. (2008: 43)

86

6 Pragmatics, Rhetoric and Semantics

6.4 Rhetoric and Argument Rhetoric is concerned with persuasion, but it is also part of the structure of discourse. All arguments must be shaped into some form of linguistic figure in order to be expressed, and like other distinctions already mentioned, that between figurative and literal use of language can be difficult to draw and to maintain. Still, there clearly is a distinction and when it is not properly recognised it can lead to bad arguments in two ways. One is that an equivocation may result when some quality is predicated of a real x and a metaphorical one. This is the case in one of the examples examined in Chapter 12, where the phrase ‘university of life’ is employed. Any argument based on a comparison of this ‘university’ and the experiences it provides with actual universities and the education they offer, is likely to fall into equivocation: the arguer has been seduced by the metaphor into thinking that the two are really types of the same thing, and can be compared like for like. Of course, there is no reason not to contrast the advantages of life experience with those of formal education, but they should not be considered as near equivalents. The second danger is in taking the metaphor too literally and extending it too far. In his election campaign of 1996, Bill Clinton referred often to the task of building a metaphorical bridge to the future (see Hinton and Budzy´nska-Daca 2019). Such a metaphor can be extended: we can ask if his bridge is wide enough to take everyone across, for instance; but there is a limit, if one were to argue against his policies on the basis that a tunnel would be better for the surrounding scenery or that his bridge was open to attack by enemy aircraft, then he would quite reasonably respond that it’s not a real bridge, and such concerns are not relevant. This is not the subtlest example, but the quiet workings of repeated figurative language on our minds and our perception of the world should not be under-estimated. There is no reason to ban metaphorical expressions from arguments, and several recent studies such as Oswald and Rihs (2014) and Macagno and Zavatta (2014) have joined Aristotle1 in defending and examining their use, so there will be no fallacy of metaphor or fallacy of figurativeness included in the list of fallacies of language in Chapter 8. There is, however, very good reason to think that metaphors, especially those so ingrained into normal language as to creep clandestinely into sentences quite unbeknownst to the speaker, can lead to all the major types of linguistic error which are sought for in the evaluation scheme presented in Part IV of this work. They can lead to ambiguity when the metaphorical thing and the real thing are taken to be the same, they can lead to evaluative implication loading when the metaphor makes a positive or negative association, and they may well lead to errors of conceptualisation, such as reification: knowledge may be your weapon, but it is not, in fact, an actual implement of war, and the metaphor, just like any analogy,2 will break down at some point. Macagno and Zavatta note that: ‘a metaphor becomes relevant to the extent that one of the possible predicates that characterize it can be used to support the presumed communicative purpose of the metaphorical utterance’ (2014: 484). When 1 See 2 See

Parson and Ziegelmueller (2003) for a discussion of Aristotle’s views in the Rhetoric. D. Gentner et al. (2001) for a discussion of the relationship between the two.

6.4 Rhetoric and Argument

87

analysing arguments featuring metaphors, it is the relevance of that characterisation which must be determined. In spite of the dangers, Scott Jacobs is right when he says: ‘While a case can be made that all fallacies involve rhetorical strategy (at least at the levels of arguer responsibility and critical judgment), we should not also conclude all rhetorical strategies involve fallacies’ (2006: 422). Any scheme for the assessment of arguments must be able to unpack the rhetorical baggage with which they may have been laden to get at the reasoning within, and not simply discard those which are imperfectly neutral or overly ornate in their expression. That last sentence, however, makes certain assumptions about how the term rhetoric is to be understood and how it is related to argument; assumptions which may not be warranted. Christian Plantin (2009) provides an excellent discussion of the argument/rhetoric divide, showing how it is based on an earlier jettisoning of a good deal of the meaning of rhetoric, leaving it as a synonym of eloquence, often mistrusted and scorned. He points out that: ‘Traditionally, there is no theoretical opposition but a complementarity between the argumentative stage, inventio, and the linguistic one, elocutio’ (2009: 327), and that the move towards understanding rhetoric as merely the elocutio did not occur until the sixteenth century. This conception is most famously and strongly reinforced by John Locke, who saw rhetoric as the enemy of good argument: if we would speak of things as they are, we must allow that all the art of rhetoric, besides order and clearness; all the artificial and figurative application of words eloquence hath invented, are for nothing else but to insinuate wrong ideas, move the passions, and thereby mislead the judgment; and so indeed are perfect cheats. (Locke 1690, III, X, 34 [1975: 508])

This appears to rule out a great deal of natural language from use in science and philosophy, and leaves us wondering just what it is that Locke considers ‘artificial’. Also, as Plantin points out, it assumes an understanding of figurative speech which ‘is grounded in the concept of a possible choice between two strings of signifiers to express the same semantic content or state of the world’ (2009: 329), that is, that there is always a straightforward way of saying things and that the use of a figure is a choice made in order to embellish or obscure the content. As a result: ‘Ornaments are worse than fallacies; they are the mask of fallacies’ (2009: 330). This conception of rhetoric has been challenged and to a large extent discarded in more recent writing. Lloyd Bitzer introduced the idea of the ‘Rhetorical Situation’ in a famous paper with that title. He sees rhetoric as the product of a certain set of circumstances and a way of effecting necessary change: rhetoric is a mode of altering reality, not by the direct application of energy to objects, but by the creation of discourse which changes reality through the mediation of thought and action. The rhetor alters reality by bringing into existence a discourse of such a character that the audience, in thought and action, is so engaged that it becomes mediator of change. In this sense rhetoric is always persuasive. (Bitzer 1968: 4)

There is no sense of rhetoric as embellishment here, indeed, there is no mention of figures. Rhetoric becomes a way of responding to a situation which requires a response. This view was rejected by Richard Vatz, who saw things rather in reverse:

88

6 Pragmatics, Rhetoric and Semantics

‘Bitzer argues that the nature of the context determines the rhetoric. But one never runs out of context. One never runs out of facts to describe a situation […] The very choice of what facts or events are relevant is a matter of pure arbitration’ (1973: 156– 157). From Vatz’s perspective, it is the rhetoric which creates the situation, rather than responding to it: ‘To the audience, events become meaningful only through their linguistic depiction’ (1973: 157). Rhetoric, then, is the creator of reality, a far remove from Locke’s conception of rhetoric as an obstacle to the discussion of what is real. Such a definition of rhetoric, without any reference to eloquence, is also found in what Plantin calls Ducrot’s ‘intuition of an argument as the ordinary capacity of a sentence to re-frame the following discourse and, more broadly, the dialogal context’ (2009: 326). All of which is to move some distance from how the word has traditionally been employed. A more common approach is to divide rhetoric into the argumentative and the ornamental: while Locke’s criticism may apply to the one, the other facets of the ancient art are preserved in the other. This is the distinction used by Perelman and Olbrechts-Tyteca in their treatment of rhetorical figures, in The New Rhetoric. They note that: ‘If the argumentative role of figures is disregarded, their study will soon seem to be a useless pastime’ (1969: 167) and state that their interest lies in ‘how and in what respects the use of particular figures is explained by the requirements of argumentation’ (1969: 168). This leads to a division between argumentative figures and figures of style; it is, however, a division which cannot be made in advance of the speech’s being made: We consider a figure to be argumentative if it brings about a change of perspective, and its use seems normal in relation to this new situation. If, on the other hand, the speech does not bring about the adherence of the hearer to this argumentative form, the figure will be considered an embellishment, a figure of style. (1969: 169)

This appears to make the nature of the figure dependent on the reception it receives from the audience. It also establishes a clear hierarchy: ‘a figure which has failed in its argumentative effect will fall to the level of a stylistic figure’ (1969: 170). Plantin is uneasy about this: since we cannot access the reception of the figure, we cannot classify it. We might then fall back on the intention of the speaker, but this does not fit with the idea that a failed figure becomes a merely stylistic one. What we can say for sure is that the authors of The New Rhetoric were interested in figures for their argumentative effect, ‘what each of them contributes to argumentation’ (1969: 171), and this is how they study them, acknowledging that the same figure may be used to different effect at different times. Plantin suggests that this is best expressed by defining the use of figures of speech as a discourse strategy: ‘a discourse strategy that has nothing to do with decoration or embellishment: these clusters of figures are the specific manifestation of the argumentative function of language’ (2009: 334). He goes on to conclude with a statement I am happy to endorse and one which expresses at once the importance of rhetoric in our understanding of discourse, and of language in our understanding of argument:

6.4 Rhetoric and Argument

89

Argumentation is necessarily embodied in language, and ordinary argumentation in ordinary language. So, a theory of argumentation is necessarily embodied in a theory of language and, more specifically in a theory of discourse. The rhetorical theory of figure is the traditional theory of discourse, and not a theory of any fallacious “supplement”. (Plantin 2009: 334)

6.5 Disputes Merely Verbal The debate over verbal disputes perhaps lies more in the realms of philosophy of language than linguistics, and is not at the heart of any theories of argumentation. However, that may be simply a result of the lack of attention paid to semantics in the field, which this book seeks to redress. Verbal disputes are usually contrasted with substantive disputes, and come down to a disagreement over the meaning of certain terms, although, at first sight, they often appear to be of a more fundamental nature. In certain fields, of course, the meaning of a term may be a very significant subject for a dispute, contract law, for instance, or, indeed, linguistics. Such cases allow us to make a distinction between verbal disputes and ‘merely’ verbal disputes (Chalmers 2011). If I say that your blouse is white and you say, no, it’s cream, then we can both go along happily together without ever bothering to even try to resolve the dispute, since nothing rests on it. If, on the other hand, you are about to take to the court at Wimbledon and an official claims that your clothing is cream, and not white, then you are in trouble, and must either convince the official or change your kit. This distinction isn’t recognised universally: for C.S.I Jenkins ‘Merely verbal disputes, then, are ones in which the dispute arises only in virtue of the parties’ divergent uses of language’ (Jenkins 2014: 20), with no mention made of whether or not something rests on that divergent use. The difference here is pleasingly ironic: Jenkins uses the word ‘merely’ in a neutral sense to mean perhaps ‘no more than’, while Chalmers removes it from non-trivial disputes since it has a pejorative sense of ‘not important’: there appears to be a small, merely verbal, dispute over the merely in merely verbal disputes. Since such disputes apparently arise due to differing beliefs over the meaning and extension of certain language items, it ought to be possible to resolve them by paying attention to those differences: where that resolution is not possible, a substantive dispute has been exposed. David Chalmers suggests that verbal analysis of disputes is a good way to make progress: firstly, ‘the diagnosis of verbal disputes has the potential to serve as a sort of universal acid in philosophical discussion, either dissolving disagreements or boiling them down to the fundamental disagreements on which they turn’ and also ‘Reflection on the existence and nature of verbal disputes can reveal something about the nature of concepts, language, and meaning’ (2011: 517). The suggested method is the elimination technique, rephrasing the dispute with a restricted vocabulary, avoiding the word which is the basis of the problem. This could certainly help in some instances. Suppose we disagree over whether a friend has lied, although we both agree that she has spoken and not told the truth. We rephrase our positions avoiding the verb ‘lie’. I may say, ‘Jane has uttered a falsehood’ as

90

6 Pragmatics, Rhetoric and Semantics

I understand this to be what lying means, your response may be to accept that she has, but to say ‘Jane did not knowingly utter a falsehood’ or ‘Jane did not sincerely utter a falsehood’. If I disagree with your statements, then it seems we do have a substantive dispute, but if I agree, our dispute has vanished and we have discovered that my understanding of ‘to lie’ is rather simplistic, while yours is more nuanced. If we were to check with other competent speakers we would also find that your understanding of the meaning was the generally accepted one and I should have to update my own beliefs about the appropriate use of the word. This dispute was not trivial, since an accusation of lying is always a serious thing, but could easily be cleared up. Other examples are not so simple to deal with. Chalmers offers the case of disagreement over Pluto’s being a planet, which is not easy to restate without the word planet, and it is hard to see how we could persuade the Wimbledon official to accept your tennis shirt without using the words ‘white’ or ‘cream’. This may be attributable to ‘vocabulary exhaustion’ which Chalmers notes is rare in a language like English: he also, more controversially suggests that ‘these exceptions will arise only for terms expressing especially primitive concepts’ (2011: 529). Such terms he later refers to as ‘bedrock terms’ which express ‘bedrock concepts’, giving the concept of ‘ought’ as a paradigmatic example. There are interesting parallels to draw here. These ‘bedrock’ terms and concepts seem to be reflected in Robert Fogelin’s argumentation category of ‘deep disagreements’ (Fogelin 2005). Just as argument requires a certain level of agreement on fundamental practices of argumentation, without which it simply hits a wall and stops, so there are certain terms which are not susceptible to further analysis and so it becomes impossible to say whether disagreement over their meaning is verbal or not. It may be that there is a sort of ‘deep disagreement’ about the usage of the words involved, or some more substantive dispute, but there is no way to find out and advance. Opinions will differ in both cases about possibilities for finding a way around the apparent obstructions. Secondly, although Chalmers doesn’t mention him, there is some reflection of G.E. Moore’s natural and non-natural objects (1903), where the natural are not capable of definition in terms of the non-natural and to attempt to do so is to commit the naturalistic fallacy (see Chapter 9), since natural terms such as ‘good’ are, by their nature, unanalysable. Chalmers himself sees a parallel with Carnap and the logical empiricist attempt to clarify philosophical problems in order to resolve them. I have given the Chalmers account here, but the discussion of verbal disputes is far from settled. Chalmers himself was reacting to work by Eli Hirsch (2005) and Theodore Sider (2009), and there has been a good deal of further discussion since. Jenkins (2014) uses the idea of prima facie disputes, which may or not contain real disputes, in his paper; Balcerak Jackson (2014) sees the verbalness of disputes as a discourse defect, but does not think it necessarily implies non-substantiveness; Rott (2015) analyses disputants ‘fact profiles’ and ‘meaning functions’, where a difference in the first means a substantive dispute, but if there is only a difference in the second, it is merely verbal. He also mentions, as do several others, that we should also be aware of merely verbal agreements. Almotahari (2019) rejects the ‘semantic

6.5 Disputes Merely Verbal

91

deflationism’ brought on by categorising philosophical questions as merely verbal, and Vermeulen (2018) brings us full circle by arguing for a pragmatic understanding of verbal disputes, invoking the Gricean distinction between utterance meaning and speaker meaning. Her paper suggests that verbal disputes can be caused by syntactic ambiguity and conversational implicature, making it a more strictly linguistics-based account, and a very interesting one for more linguistically-minded argumentation scholars. The papers cited here are just a selection, designed to show that this one type of dispute is itself the subject of much dispute, but also that there is a good deal here to interest argumentation scholars, from questions of what is a real dispute to the role of ambiguity and implicature in creating such situations. Of greatest interest at this moment and to the purposes of this work, however, are certain points which can be of relevance to the understanding of linguistic fallacies and the analysis of language as part of the assessment scheme with which I conclude. The first has to do with definitions. Chalmers points out that many philosophical questions are of the nature ‘what is x?’ where x may be knowledge, justice, meaning, and so on. Such debates are common and important in philosophy, but both likely to be in fact verbal disputes, where it is the meaning of the word which is at issue, not something more fundamental about the concept, and likely to lead disputants into one of the definitional fallacies described in Chapters 8 and 9. Arguments which include definitional fallacies are difficult to salvage through rephrasing, but the elimination method outlined here suggests that they might avoid outright rejection if the entire debate is rephrased without the contentious word. This is particularly important since a large, though rarely emphasised, part of the aims of argumentation theory is to find ways to allow for the continuation of debates which seem to be running out of road in which to establish resolution. Just as critical questions to argument schemes can serve to evaluate arguments but can also serve to extend them, so rephrasing and reconstructing arguments can work, not only to see that arguments are properly expressed, but to better clarify the crux of the disagreement and expose the differences upon which it rests. The second point is found in the question of conceptual analysis. The content of certain concepts is a common cause for disagreement in philosophy, and academic debate generally, but the realisation that disputes are common about the exact meaning of conceptual words ought to push us towards considering not what a concept is, but what role we want it to play and an acceptance that the same word/concept can have different roles. This is something I have been at pains to acknowledge in my own definitions of the vital, but variously understood concepts of ‘reasoning’, and ‘argumentation’. It is also a point which is of relevance in the consideration of both the definitional and conceptual fallacies which I describe throughout the second part of this book. Any theory of meaning which is based in a general principle of ‘meaning as use’, must conceive of words in terms of the roles they can play, rather than as symbols of a ‘real’ something which philosophers are tasked with discovering. The relationship between verbal disputes and fallacy theory is also worth exploring. When two parties are engaged in a verbal dispute, we might say that they have misunderstood the nature of language and committed the fallacies I mentioned

92

6 Pragmatics, Rhetoric and Semantics

in the last paragraph; but that would apply mainly to philosophical debates. In more mundane disagreements we could characterise the situation in two ways: either the response to the original standpoint is irrelevant and the argument is being pursued at cross-purposes, since the proponent did not mean what the antagonist has taken him to mean; or an equivocation has been committed where one of the parties is not aware that one of the uses of a term employed in the argument differed in meaning from another of the uses, and has combined them, perhaps to show the absurdity of his opponent’s position. In the former case, especially if we feel that the attack on the original argument has been made with some awareness of the possibility of a verbal dispute, we might characterise the attack as a straw man, rather than an innocent irrelevance. There might be another way of characterising what is going on when looking at natural argumentation practice. Arnulf Depperman (2000) provides a thoughtprovoking analysis of semantic shifts within arguments. Depperman looks at cases where there is no equivocation on a term, but where the semantics of a given term, often a ‘keyword’ in the debate, change and develop over its course. Rather than characterising this as a flaw in the discourse, it is seen as a normal part of the flow of argumentation. In his paper, he analyses a debate in which differing conceptions of ‘freedom’ come to be a key issue, and notes how they are moulded and developed throughout the discourse: As the instances of ‘freedom’ show, speakers actively shape the meaning of words with respect to their context of use. They do this by practices of what I would call ‘local semantic elaboration’: by explicating and exemplifying the semantics of a word, by contrasting it with other words or by establishing relations of class-inclusion, implication or synonymy. (Depperman 2000: 23)

Since this ‘local elaboration’ is common, particularly when there is a ‘verbal’ element to the dispute, semantic shifts are inevitable, but rarely lead to outright contradictions. Depperman warns against the ‘logical semantics’ which regards such changes in meaning as a route to fallacy and which he sees as prevalent in argumentation theory, and encourages a more positive view of semantic shifts. Apart from their role in precisification of meaning, there are two reasons for this: firstly, when two disputants have differing understandings of a word, they may be talking at cross purposes, but the difference may retain what he calls ‘the coherence of a confrontation between two global positions’ where opposing ‘parties interpret local moves in terms of a global positional confrontation’ (2000: 27). In the philosophical debates discussed in the literature on verbal disputes, such global differences are easier to spot since disputants generally have names for each of the positions they take, and often readily identify themselves as ‘something-ist’ before debate even begins. In the less rarefied atmosphere of everyday debate, it may only be through the refinement of one’s understanding of local terms that one’s global position becomes obvious. Secondly, Depperman cites the phenomenon of preference for disagreement (Bilmes 1991), where the structure of argument discourse favours conflict over agreement, and claims: ‘Together with higher order interpretation, this general preference for disagreement itself lends a coherent structure to the debate as a global positional

6.5 Disputes Merely Verbal

93

confrontation’; all of which explains why ‘participants don’t seem to care about obvious contradictions that result from divergent semantics’ (2000: 28). The conclusion which Depperman draws and which is, I think, worth noting and keeping in mind as we move into Part III in the next chapter, is that ‘phenomena like semantic shifts can be coherent, functional and often unproblematic for discussants. Argumentation analysis therefore should not prematurely condemn such processes as defective because of their dissociative impacts on argument structure’ (2000: 29). One final reason why argumentation theorists might gain from looking at the category of verbal disputes is the, it seems to me, quite widespread prevalence of retorts to arguments along the lines of ‘that’s just semantics’ or ‘you’re splitting hairs’ as though the precise meanings of words in arguments were not of any real concern. Some examples of this can be quite fascinating. This is from an interview with a Nigerian physician: ‘There is misinformation going on. There is recreational and medicinal marijuana, those are just semantics. Marijuana is marijuana’ (Okundia and Efunla 2019). Intriguingly, the doctor goes on to describe quite clearly the difference in use of the two terms, but his point seems to be that recreational marijuana is a misnomer. A clearer use of the phrase comes in a second example where foreign students in the UK accused of cheating in an English language test had visas revoked and had to leave the country. A report quotes a government spokeswoman and gives a response: ‘Immigration Minister Caroline Nokes defended the government’s actions as “measured and proportionate” and said that the students were removed, but not deported from the country. This is just semantics for those affected’ (Sharma 2019). In this case the question of whether the process is best termed ‘removal’ or ‘deportation’ is a merely verbal one, and is of no consequence to those who have had to leave. The complaint that it is just semantics is actually an accusation of irrelevant argument and, from the writer’s point of view, seems justified, although in a legal context, this would not be ‘just semantics’, but might determine whether the removals were in line with the law or not. If the ‘it’s just semantics’ response is as popular an argument move as the number of websites discussing it would suggest, then argumentation theory ought to be ready with a description of when it is justified and when it is a fallacious accusation. That will require a proper description of verbal disputes in the argumentation literature integrating them with more general theoretical structures.

Part III

Error

Language disguises thought Ludwig Wittgenstein, Tractatus, 4.002

Chapter 7

Fallacies

The concept of fallacy is of such fundamental importance to the assessment of arguments because it presupposes that the errors made within them are to some degree generalizable. Without fallacies, we have only individual arguments which have gone wrong, even though they may have perfectly achieved that which they were designed to do, that is, typically, to persuade somebody of the truth of some proposition. If every instance had to be considered individually, and the case made as to why the argument should not be admitted, without reference to any standard patterns of rejected structures, the field would be in a sorry state. The good and the bad must be separated with regards to something, and it is rather easier to recognise and employ patterns of bad argument than good. Fallacies, then, are not only types which can be identified, collected and studied, they can also be seen as breaches of some normative standard—yes, your argument may have persuaded millions of people to vote for you or buy your product, but it was still a bad argument, which, in some sense, shouldn’t have been accepted. Scholars may disagree, and they do disagree, on what fallacies are, but everyone knows for certain that they are a bad thing: winning an argument through deliberately fallacious reasoning is simply unfair and deceitful, it is poor form, and has been from the Sophists to the modern purveyors of ‘fake news’. Part III of this book, will not try to set out a complete history of thought about fallacies or to analyse the many theories and definitions currently available, although a little introduction will be necessary and some will be mentioned as the discussion proceeds. The focus here is on two elements which ultimately merge into one main concern: how fallacies can be understood within my own theory of argument as described in Part II above, and, in the following chapter, how fallacies of language have generally been treated. These two strands are then brought together quite naturally in certain conclusions as to the nature of linguistic fallacies in my system, and how they relate to the general view, if such a thing can be established.

© Springer Nature Switzerland AG 2021 M. Hinton, Evaluating the Language of Argument, Argumentation Library 37, https://doi.org/10.1007/978-3-030-61694-6_7

97

98

7 Fallacies

7.1 A Standard Definition Certainly, the concept of fallacy can be understood in many ways. While Hamblin’s (1970) insistence that the standard definition of a fallacy was concerned merely with questions of validity may have been thoroughly debunked by Hansen’s (2002) exhaustive review of the earlier literature, there is no doubt that that is one way in which fallacy could be understood. Indeed, one might feel able to make the claim that logical fallacies which produce invalid arguments are actually to be regarded as the true fallacies, involving, as they do, false inferences, rather than merely false or irrelevant premises. This view may not be popular in the literature of argumentation, but it is clear that formal logical fallacies are errors of a very different nature to the informal fallacies. Fallacies, in fact, have been understood in many different ways and the only way to embrace all of those errors which are usually included in such lists is to take a very broad understanding of the concept indeed. Hamblin’s famous characterisation of traditional thinking runs: ‘A fallacious argument, as almost every account from Aristotle onwards tells you, is one that seems to be valid but is not so’ (Hamblin 1970: 12), which tells us that, supposedly, fallacies are arguments, that they seem valid, and that they are not valid. Hamblin, quite correctly, points out that this definition does not fit many of the traditional fallacies and is not, therefore, fit for purpose. Hansen’s later article exposed Hamblin’s dismissal of this definition as a fine example of a Straw Man: of the more than twenty works investigated, from Aristotle through to publications appearing in the 1960s, almost all did define fallacies as arguments, a narrow majority agreed that they should look like good arguments, and almost none actually considered invalidity to be a defining feature (Hansen 2002). Indeed, Whately not only defines fallacy, but defines it as he believes it is ‘commonly understood’ to be ‘any unsound mode of arguing, which appears to demand our conviction, and to be decisive of the question in hand, when in fairness it is not’ (1857: 101). None of which necessarily negates Hamblin’s wider point that ‘we have no theory of fallacy at all, in the sense in which we have theories of correct reasoning or inference’ (1970: 11), but it does illustrate that the concept of fallacy has long been understood to be broader than a mere question of validity. It was, in fact, van Eemeren and Grootendorst who introduced and popularised the term ‘standard definition of fallacy’ and its lack of fit with reality was useful to them in introducing a new approach to argumentation in general, including to fallacies. They note that: ‘In our opinion, a satisfactory and more-encompassing alternative to the standard approach can only be fully developed if the preoccupation with the logic of arguments is wholly abandoned’ (1992b: 152); and indeed, by far the neatest theory of fallacy is that offered by the Pragma-dialecticians. The view that anything which breaks the rules of argument, as laid out by Pragma-dialectics, is a fallacious move has a number of advantages. Firstly, it makes it clear what is a fallacy and what is not, by showing exactly what an argument move needs to do in order to qualify as one. Secondly, it puts no restrictions on the nature of fallacies since it is their effect, the prevention of the resolution of the critical discussion, rather than any element of their internal structure, which leads to their being so-defined. This clarity, though,

7.1 A Standard Definition

99

comes at a price: many types of behaviour within the argument setting which break the rules, such as refusing to defend one’s own assertions when they are questioned, are simply not the sort of thing which other theorists of fallacy have been interested in. That is largely because one thing most theorists have agreed upon, as Hansen’s review makes clear, is that fallacies are arguments, while it is clear that one does not have to offer an argument to break the rules of the critical discussion. This apparent shift in the meaning of ‘fallacy’ might well be considered an example of Leonard Nelson’s ‘philosophical fallacy’ (2016), of which much more will be said in the coming chapters. Some other objections to the pragma-dialectic conception of fallacy are listed by Tindale (1996). One of these, anticipated by Woods (1988), is that either there are only ten fallacies to match the ten rules, or there are more fallacies which break the rules in different ways. This latter thought, however, seems to suggest that there is something fallacious about those fallacies which is not simply captured by the idea that they break the rule. The traditional fallacies can be listed as constituting the breaking of particular rules, but then they remain understood as fallacious forms in their own right. I’m not sure this is an important criticism. An individual argument form, or move, has its own recognisable structure, and may be given a name, but, on this account, it only becomes a fallacy by virtue of its breaking a rule. Thus, the use of many traditional names for fallacies doesn’t change the fact that there are only as many ways to be fallacious as there are rules. Tindale, too, is more troubled by the fact that what is acceptable in a critical discussion is decided solely by the participants: ‘this attitude seems to eschew any kind of objective criteria for judging fallaciousness’ (1996: 24). The only normative standard introduced which goes beyond the opinions of those involved is that argument moves should facilitate the reaching of an agreed conclusion. This means that utterly ridiculous arguments are acceptable if the antagonist and protagonist agree to accept them, but, more importantly, that an argument can be rejected by one of them as unsuitable if it harms their cause. Tindale quotes van Eemeren and Grootendorst stating that the participants must ‘decide about whether or not to allow this type of argument in their discussion. If they cannot reach an agreement, the protagonist should not use it’ (van Eemeren and Grootendorst 1992a: 66). At which Tindale notes ‘the upshot of this will be that perfectly good arguments will be dismissed because they do not contribute to resolving disputes’ (1996: 26). The flaw in Tindale’s reasoning is obvious—he is begging the question by using a concept of good argument to determine what is a good argument; his claim is that contribution to resolving disputes is not the definition of a good argument because some good arguments, on some other definition, don’t help to resolve disputes. To be fair to Tindale, the excerpts which he cites seem to show that the pragma-dialecticians are also reluctant to give up the notion of a ‘good argument’ defined rather more objectively, although there is a clear conflict in their discussion. John Woods presents an intriguing argument which contains both a more realistic conception of what the standard view of fallacies has been, and a criticism of the pragma-dialectical re-purposing of the term. His examination of the topic, however, ultimately leads to a rehabilitation of pragma-dialectics and a great deal of doubt’s

100

7 Fallacies

being thrown on the traditional thinking which he might usually be taken to represent. Woods (2007) describes the standard approach as the EAUI conception. That is, to be fallacious, a pattern of reasoning must be erroneous, attractive, universal, and incorrigible. This view he claims to be widespread, meaning that van Eemeren and Grootendorst’s use of the term is guilty ‘of substituting a stipulated definition of fallacy for the traditional conception’ (2007: 74). Woods then goes on to examine the literature on the best-known fallacies to see what is thought to make them fallacious and comes to the conclusion that there is a ‘striking consensus to the effect that the mistake embedded in most of these fallacies is either the error of invalidity or the error of inductive weakness’ (2007: 79). The question then becomes whether inductive weakness constitutes an error. Woods is doubtful; in cases such as arguments from ignorance, he suggests that, given the resources we have available to us, material and cognitive, it is not an error to employ or accept arguments which are not formally valid. At the same time, argument forms which are errors are not attractive, universal or incorrigible. This leads to the tentative conclusion that, as far as the traditional EAUI conception is concerned, ‘there are no such fallacies’ (2007: 87). Hence, there are either no fallacies at all or a different conception from the traditional EAUI must be found. Woods then admits: ‘if, as I now conjecture, the traditional concept of fallacy is indeed empty, it is much harder to persist with one’s dissatisfactions with the Amsterdam School’ (2007: 88). That is not to say he accepts that school’s fallacy theory, but it does mean that he concedes the need to either replace the standard understanding with something else, or to embrace the title of his essay—The Concept of Fallacy is Empty. It seems, then, that the idea that a fallacy is a false move in an argumentative discourse, rather than simply a falsely constructed argument, is worth holding on to, if we wish to maintain the notion at all. However, even if we accept the list of rules as complete and a fallacy as a breaking of the rules, that is not the whole story. There is still a need to examine why particular patterns of reasoning and particular forms of expression are likely to be considered breaches of good practice in a critical discussion or indeed any other argumentative environment. We should also bear in mind that this idea is not identical with the traditional conception of fallacy as argument, so that we don’t find ourselves equivocating or arguing at cross purposes because of the difference. Much of the discussion below is of theorists who work or worked within the traditional approach to fallacy, and their use of the term should be understood accordingly.

7.2 Categorisation A favourite occupation which both unites and further divides theorists is the splitting of fallacies into categories; this is a game I play myself a little further below. The characteristic features of these divisions are that they are at once always authoritative and never the same. Aristotle, as we shall see in detail in the following chapter, began this practice by dividing his sophisms into those connected with language and those

7.2 Categorisation

101

not so-connected. Richard Whately suggested that this distinction had been kept, but ‘never grounded on any principle’ (1857: 104) to which theorists had actually adhered. He is at pains to show how his distinction into Logical and Non-logical fallacies would ‘keep clear of all this indistinctness and perplexity’ (1857: 104), something which those who have followed have achieved to greater or lesser extents. He allows a further division of the first category into Purely-logical and Semi-logical, and the second, which he also calls Material fallacies, into ‘Premiss unduly assumed’ and ‘Conclusion irrelevant’ (1857: 108), categories found under similar names in the later authors below. One other point of interest is that in the table he produces showing this division and placing various fallacious argument types under the given headings, Whately makes no mention of fallacies of language. He does, however, discuss the effects of language on the detection of fallacies, stating that a fallacy is more likely to succeed in persuasion ‘the more it is obscured and disguised by obliquity and complexity of expression’ (1857: 112). This notion, that imprecise language hides weaknesses in argument, rather than actually constituting one itself is deserving of some attention and will be discussed in the following chapter. John Stuart Mill also takes a logical approach, though it is to some extent more practical, and makes a basic distinction between fallacies of Simple Inspection—a priori fallacies—and fallacies of Inference. Of this latter type, there are four subtypes: of Observation, Generalisation, Ratiocination, and Confusion, of which the first two are inductive and the others deductive fallacies. Within the a priori category Mill includes ‘cases in which a proposition is believed and held for true, literally without any extrinsic evidence’ and ‘those more frequent cases in which simple inspection creates a presumption in favor of a proposition’ while those which involve inference ‘must be subdivided according to the nature of the apparent evidence from which the conclusions are drawn’ (Mill 1882: 903). For Mill, as an empiricist, the key concept in reasoning is evidence, and the key distinctions among errors of reasoning are the absence, presence and nature of any evidence upon which they are based. Moving into more recent approaches, Hurley has fallacies of Relevance, of Weak Induction, of Presumption, of Ambiguity, and intriguingly, of Grammatical Analogy. This final category consists of just two fallacies: those of composition and of division, understood in the modern, not the Aristotelean sense (see next chapter). I find the explanation that: ‘Arguments that commit these fallacies are grammatically analogous to other arguments that are good in every respect’ (Hurley 2000: 156) to be deeply enigmatic. Doesn’t every bad argument have basically the same grammatical form as every good one? Sadly, there is no more discussion of how this distinction was made or why it was so-christened. The much-followed logic textbook by Copi et al. ignores this unusual category and makes a fourfold distinction, identifying fallacies of Relevance, Presumption, Ambiguity, and Defective Induction; this last being less common in the general introductions found on the internet, which otherwise seem to take this work as their basis. When giving the rationale behind this division, the authors note that ‘classification of fallacies is a controversial matter in logic. There is no one correct taxonomy of fallacies’ (2014: 125) and make it clear that the main principle they have employed is ease of understanding for the student coupled with inclusion of all common fallacies.

102

7 Fallacies

They also note that the most frequently encountered species are those of relevance. Perhaps this is not surprising: van Eemeren and Grootendorst take a somewhat jaundiced view of this category of relevance: ‘In the standard approach to fallacies there is always a category of fallacies of relevance. Its prime function seems to be that of a depository for all those fallacies for which the standard approach has no adequate solution’ (1992b: 147)! The task of classification and sub-division can be taken to new, deeper, levels too. Walton has a paper dedicated to classification within the realm of relevance. His discussion results in a diagram in which fallacies of relevance are renamed as ‘Misdirected Argumentation’ and divided into two main types ‘Wrong conclusion’ and ‘Red Herring’. What is interesting is that the examples analysed make it clear that while these two varieties of irrelevance are separate phenomena, there are cases to which both descriptions are applicable. This means that the third level of his diagram has three boxes: pure wrong conclusion, called ‘Missing the Point’; pure red herring, called ‘Petti-fogging’; and the mixed variety, ‘Diversionary’ (Walton 2004: 99). Arne Naess foreshadows the pragma-dialecticians by offering six rules to follow which he calls ‘six principles for relevant discussion’ (1966: 122). Each principle is given a name, and then a formulation in normative terms; for instance, the first principle is ‘Avoiding tendentious references to side issues’, which is formulated as: ‘One should keep to the point (even if one is aware that it harms one’s own interests to do so)’ (1966: 122). The others also call for the avoidance of the tendentious in: renderings of other people’s views, ambiguity, argument from alleged implication, first-hand report; and use of contexts. Although Naess does not explicitly do so, one could simply declare any breach of these rules to be a fallacy of relevance under that heading. What is of particular interest to this study is that principles two to six are all, in fact, principles of language use. The formulation of rule six, concerned with contexts is: ‘A matter should be presented in a neutral way in a neutral setting’ (1966: 131), and since arguments are presented linguistically, this is as much a linguistic rule as the one on ambiguity. That Naess sees the relevance of an ‘effective discussion’ to be so bound up with the way in which language is used both makes him a friend of this undertaking, and illustrates the interconnectedness of the elements of argumentation. The variety of approaches is infinite: Finocchiaro defines ‘six types of fallaciousness: formal, explanatory, presuppositional, positive, semantical, and persuasive’ (1987: 270). Whilst this is a praiseworthy attempt to consider varieties of fallaciousness, rather than named fallacies, the technical vocabulary used is likely to serve to confuse: ‘persuasive’ fallaciousness refers only to ‘the flaw of begging the question and circularity’ (1987: 270), so the name reveals little of what is meant, and ‘semantical’ is restricted in the description to cases ‘when the conclusion does not follow because the premises contain a term which has two meanings’ (1987: 269) and there is an equivocation on this term. I shall suggest below that the semantics of the words used in any given argument can lead to fallaciousness via more avenues than simple ambiguity. Johnson and Blair, on the other hand, begin their textbook with a chapter on three basic fallacies, then group the rest into fallacies of diversion, impersonation, language and meaning, and intimidation; Tindale (2007) begins with categories of diversion, and structure and language, before three chapters on ‘ad’ arguments: one

7.2 Categorisation

103

on ad hominem, one on ad verecundiam, and one of the rest, claiming that they all involve ‘appeals to some contextual factors that characterize the type of argument involved’ (2007: 104). There would appear to be, then, as many ways to categorise fallacies as there are authors willing to do so. What is important, and should become clear in my own classification below, is that, as Copi et al. noted, the challenge is not to find the correct division, but rather to organise fallacies in such a way as to allow them to be understood in the way necessary for whatever the purpose of the author or his readers may be. A little-known example of this can be found, rather surprisingly, in the last, unfinished novel of Charles Dickens ‘The Mystery of Edwin Drood’. There is in that work a character called Mr. Crisparkle who delivers a tirade against unfair debating practice at a gentlemen much given to fallacious reasonings, a Mr. Honeythunder. During this speech, four examples are offered of what Crisparkle, quite rightly, objects to as ‘platform practices’ (Dickens 1899: 212). In each of the examples there is a common element, yet they are hard to fit into one accepted category, all contain the unfair ascription of a certain belief to one’s rival. There is something of the ad hominem, something of the Straw Man, and some admixture of other irrelevancies too. What is important here is that they are grouped together not because their structure is identical, but rather because the intention behind them is; reflecting the mode of thinking about an opponent in a debate.1 The grouping of fallacies, then, is not a project which can reach a clear resolution, since some other grouping may be more relevant and more helpful for some other purpose. Walton makes the point that ‘the classification of fallacies (incorrect arguments) needs to be based on an underlying classification and analysis of argumentation schemes (correct forms of argumentation)’ (2004: 102). The problem being that despite the enormous amount of analysis that Walton himself and many others have undertaken of argumentation schemes, it isn’t clear that any good classification has come out of it (although, see the discussion of Wagemans in Chapter 11 for a very promising suggestion). In the absence of a generally accepted classification, the alternative is to take all of them individually and keep them in one, undivided collection.

7.3 Fallacies Uncontrolled One problem with the collection of fallacies is that they are so very popular: at last count, the Wikipedia list contained around 120 items,2 all listed as separate patterns of fallacious reasoning; which itself is nothing compared to the 300 collected and described by Bo Bennett, Ph.D., in his opus magnum Logically Fallacious (2018). It is also true that within particular fields fallacious reasoning used in a particular way may be named for its relevance to that field when it is actually an instance of 1 For 2 The

a full discussion see Hinton (2020a). list is available at https://en.wikipedia.org/wiki/List_of_fallacies.

104

7 Fallacies

a more general pattern. An example of this process can be found in Clark’s (1973) discussion of experiments which use language as an input and take the different linguistic items to be a non-variable factor, which he refers to as the “language-asfixed-effect fallacy”. The danger of this is real and has been widely discussed. As Raaijmakers, Schrijnemakers, and Gremmen explain: ‘Selecting language materials randomly or pseudorandomly leads to sampling variance that must be taken into account. Otherwise this variance will be confounded with the effect of the treatment variable’ (1999: 416). The point being that not all language items are the same, and, for example, it cannot be assumed that response times to particular items by one subject can be directly compared to response times of a different subject to different items. Whilst for those using statistical analyses of psychological experiments where language is used as a stimulus this may be best-remembered under the name Clark gave it, for scholars of argumentation, it is clearly an example of a generalisation fallacy, where items with some element in common, i.e. being words, are assumed to have identical other properties. Clark’s fallacy, then, is not a new discovery in itself, but his work, and that of Coleman (1964), point out how the fallacy has been lurking in a particular place unnoticed, and the name he has given it serves as a warning to researchers working in that area. This example also shows how the two commonly used meanings of the word fallacy can be entwined, and how confusion may result. While argumentation theorists use fallacy to mean a type of flawed argument, or, at least, of error in argumentation, the general populace might be more likely to use it to mean a common, but false belief. In the case above, there was a belief amongst researchers that language could be used as a fixed variable: that belief might be described as a fallacy; or we might claim that an erroneous belief led researchers to reason incorrectly about what their results showed. In the statement: ‘The major problem with this so-called “language-as-fixed-effect fallacy” is that it increases the probability of Type I errors, i.e., concluding that a treatment variable has an effect where in reality there is no such effect’ (Raaijmakers et al. 1999: 416), it isn’t clear whether the fallacy means a belief or a pattern of reasoning. The distinction between fallacious argument structures and arguments based on false beliefs is a difficult one to maintain—the gambler’s fallacy, for instance, is an argument which relies on a false premise about the probability of an event, but the general idea that luck must change, that a run of heads makes a tail more likely, which is the basis for that false premise, is itself a fallacy in the popular sense. Even the apparently obvious distinction of formal and informal fallacies is questionable. Argument schemes, many of which are potentially fallacious, but can also serve as important patterns of ‘presumptive’ reasoning, can usually be written out in either a deductive or inductive way. In his earlier book, Walton’s ‘appeal to expert opinion’ for example, has as its conclusion ‘A [the assertion] may plausibly be taken to be true (false)’ (Walton 1997: 223), but in a later collection simply ‘A is true’ (Walton et al. 2008: 310). With the first conclusion, the argument continues to be valid, even when A is, in fact, false. The more straightforward conclusion ‘A is true’, on the other hand, makes the argument inductive and in the strictest sense commits a formal fallacy, whether or not it is actually true.

7.3 Fallacies Uncontrolled

105

In Walton’s extensive work on such schemes the presumptive nature is sometimes written into the argument and sometimes alluded to in the surrounding text, an inconsistency which carries over into, and runs through, the exhaustive compilation of Walton et al. (2008). Whether or not an appeal to an expert or a person in a position to know is a formal or informal error when that person happens to be wrong appears to be a question merely of how one sets out the argument in writing. It may be, inductively, reasonable to believe the conclusion of an argument, that p, or the conclusion of the argument may be that it is, deductively, reasonable to believe that p.

7.4 Fallacies in the Present Framework of Argument General thoughts on the nature of fallacy and the state of fallacy theory can be left to be expanded upon elsewhere and elsewhen. The conception of argument laid out above in Chapter 4 left three clear areas in which individual arguments could go wrong, three points at which the arguer may make a fallacious move. These three were defined as the process, the reasoning, and the expression in language. I shall also claim that they are not to be considered as parallel distinctions: rather that they form a hierarchy, which I describe below. The process, let us remind ourselves, refers to both the situation in which the argument is being delivered and the established or accepted practices of argument associated with it. Examples of commonly described fallacies committed at this stage would be Locke’s ad verecundiam ([1690] 1975), where a statement may not be challenged because of the status of its author; the Straw Man, where an opponent is misrepresented; and any type of irrelevant argument, be it Circular reasoning, which fails to move the debate forward; ad hominem,3 which fails to address the issue at hand and concentrates on the arguer; or ad misericordiam, again, appealing to factors outside the matter of the debate. These types of errors are the particular concern of pragma-dialectics: of the 10 basic rules laid out in the Handbook of Argumentation theory (van Eemeren et al. 2014: 550–551), 7 refer to what might be termed procedural matters; rules 7 and 8 consider the actual structure of arguments, and rule 10 the language. I would also suggest that they are found at the lowest level of poor argumentation. Many of the violations of those 7 rules are little more than ill-mannered behaviour on the part of the rule-breaker; certainly, ad hominem and Straw Man arguments are associated with those who wish to discredit opponents rather than explicate good arguments of their own in order to reach an accepted conclusion. Sometimes these errors may be the result of a simple lack of training or experience in the giving of arguments, othertimes they may be a pernicious tactic of the underhand. Either way, they belong to the lowest level of sophistication in argument, they often involve a refusal to enter seriously or sincerely into a debate at 3 For

an interesting discussion of ad hominem arguments, including their non-fallacious use see Metcalf (2005), or, more recently, Martini (2018).

106

7 Fallacies

all, and can scarcely be thought to reveal errors in the thinking of an individual on a given topic, since, by their nature, they refuse to engage properly with the relevant thought. Reasoning was understood to be the internal, pre-linguistic process of inference making. Obviously, errors occurring at this stage would include all the strictly logical, formal fallacies, such as denying the antecedent and affirming the consequent; these are the moves covered by rule 7 of the pragma-dialectics. The human mind, however, is not restricted to purely deductive thinking, and is just as likely to reason on the balance of probabilities, making use of imperfect evidence, background knowledge and beliefs, and experience. When such reasoning is employed, the identification of the error, or fallacy, becomes more subjective. Take the fallacy of anecdotal evidence: this is a pattern where a conclusion is reached on the basis of insufficient evidence, usually because of a lack of understanding of the nature of good evidential reasoning—a scientist who has data, but too few to support his claims would not be accused of relying on ‘anecdote’, although the flaw in his argument would be of a similar nature—but who is to say how much evidence is enough? What is the borderline between anecdote and evidence? A family tale of one great uncle who lived to a hundred despite drinking a bottle of whisky a day is clearly anecdote, and a scientific study of 10,000 heavy drinkers is evidence, but there may be many cases in between. It is clear, then, that the three elements of argument do not stand alone, without relation. It is the nature of the process which authorises certain types of reasoning. The reasoning acceptable in a court of law is not the same as the reasoning acceptable to scientific journals or the reasoning relied upon by groups of friends in relaxed conversation. The categories of assessment may be identical, but the standards of permissible inference are not. This is why a number of the presumptive argument patterns in common use, such as argument from expert opinion, may sometimes be considered fallacious and at other times acceptable and fair: different situations call for different degrees of proof. The third element of argumentation is the expression of that reasoning in language; and the third type of fallacy, therefore, is the fallacies of expression. This is a particularly problematic stage for even the best trained maker of inferences with the sincerest and best intentions. The inference, and the stages which have allowed the reasoner to reach it, must be coded into words which he believes will be understood as conveying the meaning he intends, and, just as importantly, they must be so understood by the intended audience. A statement which is vague and imprecise may be considered a failure of the first stage, whereas an inability to grasp certain concepts or ignorance of certain vocabulary causes a breakdown at the second stage. It is not normal practice to ascribe fallacies to the receiver rather than the maker of arguments: more likely it would be the response of that receiver which would be considered fallacious—it may well be guilty of irrelevance, for instance, if the original argument has not been well understood. It may seem unfair to make not being understood a fallacy, as an arguer can hardly be expected to know which words his audience knows and which not, yet it is clearly incumbent on all parties not to deliberately obscure their language with esoteric words and jargon, and even to adjust their style and vocabulary range

7.4 Fallacies in the Present Framework of Argument

107

appropriately, as when talking to a child, a non-native speaker, or a person of limited education. This kind of behaviour, either deliberately or carelessly talking over the head of the audience, might well be better considered a fallacy of process than of language, however. This would presumably qualify under Rule 10 of the pragmadialectics as ‘misuing unclearness’ (van Eemeren et al. 2014: 551), but it is a rather different argumentation sin than some of the others so-listed. I have mentioned above that these three varieties of fallacy, defined by the stage at which they occur, rather than any internal characteristic, stand in an hierarchical relation to one another. There are three ways in which I think this is the case. Firstly, I mean that while there is little doubt that they all occur in all arenas of debate, the fallacies of process can be considered the lowest type, used either as a deliberate scuppering tactic to avoid serious discussion, or by the most uncouth of participants who have no understanding of the process in which they are engaged; fallacies in reasoning may also be used nefariously by the subtler Sophist, or may occur when an arguer in good faith simply makes an error of inference, through lack of training or experience, and, crucially, can be argued with and have his error explained to him; fallacies of expression, though, are the subtlest of the three: only the cleverest arguer can use them deliberately to make serious points—see the discussion of Aristotle’s examples below—and they may often go unnoticed by even the most skilled and practised thinkers. This ordering may be expressed thus: fallacies of process are a throwback to the sandpit, they occur when no real debate is intended, they can be easily removed from serious situations, such as in a court of law; fallacies of reasoning may be associated with politer forms of discussion, the undergraduate essay, the friendly bar-room chat, the naïve newspaper column, where there is serious discussion, but not at the most sophisticated level; fallacies of expression are committed by fine minds engaged in the most intellectual levels of debate; those who would be unlikely to abuse the process or infer poorly. This last claim is supported by the theory of Leonard Nelson, mentioned above. His ‘philosophical’ fallacy, which is described as being particularly characteristic of that type of discourse, is a species of equivocation, which allows the creation of false dichotomies. In his excellent introduction to the recent translation, Fernando Leal elegantly describes one step in the fallacy as: ‘to replace the meaning of whatever piece of vocabulary we are using with a very special and different concept or idea, the consequence of that concept-swapping being that one can now prove practically anything’ (2016: 5). The theory itself and the evidence that this practice is common among philosophers will be looked at in detail in Chapter 9. For now, let it be said only that the naming of linguistically-derived errors as those most subtle and most likely to befall the finest minds is an idea which has been raised before. The hierarchy is also important in the analysis of arguments. First, an argument, or a move in the argument, can be analysed to see if it is a fair part of the process; if it is not, then it is rejected without further attention. Once a move is accepted within the bounds of the process, the conclusion, or standpoint, offered can be assessed as an inference from the given premises, themselves subject to possible criticism, and accepted or rejected on that basis. Finally, arguments which are accepted as fair process, and appear to employ sound inference, can, if any doubt remains, be

108

7 Fallacies

subjected to a careful linguistic analysis. Exactly what that analysis should consist of and how it can be conducted are matters discussed in the closing chapters of this book. At this stage, however, it can be noted that this ordering looks reasonable when we consider what happens if we reverse it: it would be nonsense to carefully analyse the wording of an argument and then to later reject it because it was not an argument at all and merely an insult; similarly, to examine the inference employed before realising that the subject of the conclusion is of no relevance to the debate would be a waste of one’s time. Put simply, if an argument move is fair process, we assess its inference. If its inference is also acceptable, then, particularly if we are still not inclined to accept the conclusion or have some lingering doubts about what appears to have been proven, we subject the expression of the whole, premises and conclusion, to a precise linguistic examination. The final sense in which the hierarchy is expressed is in the importance, at least from an intellectual perspective, of developing tools to identify and expose fallacious moves in argument. Fallacies of process may be successful in achieving serious real-world impacts: politicians employ them constantly in their election campaigns, misrepresenting their opponents, failing to defend their own proposals with serious reasoning, and frequently indulging in the irrelevant, but they are easily dismissed by the serious-minded and those with some intellectual curiosity. Fallacies of inference can be more dangerous in this intellectual sense, as they can only be recognised by those with some capacity to analyse actual reasoning, rather than behaviour in an argument. Fallacies of expression, however, which I have several times described as subtler than the other varieties, and of which I shall provide examples in the chapters below, are capable of doing the most severe intellectual damage by leading astray serious political thinkers, academics and philosophers. These matters have not, I believe, received much attention in the literature: a fuller understanding of what the fallacies of expression actually are, in my system and in the thinking of others, will be necessary before the concept of the hierarchy can be fully understood.

Chapter 8

Fallacies of Language

I turn now to the relationship between my conception of fallacies of expression and some views on fallacies of language which have been described in the literature. The obvious place to begin is with Aristotle’s Sophistical Refutations. This is for two reasons: Aristotle’s work contains the first list of Sophisms or fallacies, and, unlike modern argument scholars, he finds the most important distinction between them to be that some depend on language and others do not: ‘There are two methods of refutations; one has to do with the language used, the other is unconnected with the language’ (Aristotle 1955: 13). This brief description is followed by a review of modern literature on a range of errors variously described as being fallacies of language, ending with the introduction of the novel idea, in fallacy theory at least, of the error of semantic incompatibility.

8.1 Aristotle’s List From the beginning of his discussion, Aristotle sets up an “inviolable distinction between the linguistic and the extralinguistic sources of error” (Schreiber 2003: 3). Of the 13 bad practices he identifies, 6 are of language, 7 not of language. The list of linguistic fallacies: Homonymy, Amphiboly, Form of the Expression, Accent, Composition, and Division, will be discussed a little more below; but the distinction itself comes from the fundamental position that “false reasoning is persuasive only if the victim holds a particular false presupposition about either language or the world. It is the nature of that presupposition that determines where the example of false reasoning is situated in Aristotle’s typology” (Schreiber 2003: 4). For Aristotle, then, errors in argument come from a weakness either in understanding language or in knowledge of the world. Homonymy occurs when two differing things have the same name. When this is the case any use of that name is necessarily ambiguous. Ambiguity itself, as we have seen, cannot be considered any sort of error, but ambiguous language makes © Springer Nature Switzerland AG 2021 M. Hinton, Evaluating the Language of Argument, Argumentation Library 37, https://doi.org/10.1007/978-3-030-61694-6_8

109

110

8 Fallacies of Language

an argument vulnerable to the trap of equivocation. An equivocal argument is one where the inference relies on a switch of meaning in the same word, and, as a result, the conclusion does not follow from the premises, although it appears to. Aristotle gives three examples: of the Greek verb ‘to learn’ which, as in the English, can mean to study something which one doesn’t know, like a new language, or can mean to employ an already acquired skill to examine particular cases; of the adjective ‘good’ which is understood to mean both ‘necessary’ and ‘desirable’, though not everything which is the one is also the other; and finally of descriptions which may truthfully apply to the same entity at different times, creating an apparent contradiction when some predicate is both true and not true of the same subject. The switch in meaning can occur within the premises or in the conclusion: ‘the sophistical appearance of valid reasoning is produced by a possible reading of the argument as valid. The price that one must pay for that reading is either an irrelevant conclusion or a false (nonendoxic) premise, depending upon whether the homonymy is in the conclusion or the premisses’ (Schreiber 2003: 25). It is, then, a little confusing to name this type of fallacy ‘homonymy’; since homonymy is a simple phenomenon of language which is being exploited by the fallacious arguer. Such fallacious moves are better referred to as fallacies of equivocation, and I shall argue below that this species of error is endemic at the highest levels of argument practice. Amphiboly can be most simply defined as a double meaning brought about by ambiguous syntax, rather than semantics. Uninflected languages, like English, are particularly at risk, since, outside of personal pronouns, there is nothing to distinguish grammatical subjects from objects, or even verbs from nouns in many cases, but the unreliable patterns of word order. Indeed, but for conventions of word order, practically every English sentence would be susceptible of multiple parsings, giving them an importance that foreign students of the language do not always immediately grasp. That said, those conventions do hold, and the deliberate creation of amphibolous sentences often means a suspension of those rules or an unnatural construction. Take the phrase ‘I’ll show you all there is to see there’. This could be understood as a promise to take one to all the places of interest in the area, or as an undertaking to reveal some other secrets upon arrival at a particular place. Yet that second interpretation would represent an unusual construction—one would normally say ‘when we get there’ if that were the intended meaning. Because of this, although the phenomenon of amphiboly could lead to a deceptive argument, I would suggest that in practice it is not often a source of real problems in argumentative discourse in English. In his posthumously published article on the subject, Douglas Walton also expresses such doubts: ‘one is left wondering whether amphiboly is a serious fallacy worth pursuing as a logical problem that can really cause significant trouble in realistic argumentation’ (2020: 13). Arguments about amphibolous sentences, however, are a different story, the Second Amendment of the United States constitution being a famous example— although that might also be cited as a simple case of poor sentence construction,

8.1 Aristotle’s List

111

compounded by odd punctuation, which doesn’t actually admit of any reasonable interpretation!1 This point is only emphasised by the five examples Aristotle offers in section 4 of the Sophistical refutations: three of them are of sentences which are not part of arguments, and the two which are set out within argument structures are better considered as riddles than as real fallacies: indeed, riddles generally rely on some form of ambiguity. The fourth example can be rendered into English thus: P1: Whatever you claim to be, you claim you are. P2: You claim a stone to be. Conclusion: You claim you are a stone. Such arguments may be of use to confuse an opponent and may also be difficult to refute if the ‘riddle’ is difficult to solve, but they do nothing to actually convince anyone of the truth of their conclusions. No-one will accept that he is really claiming to be a stone, when he claims that a stone exists. The ‘argument’ highlights an amphibolous construction with ‘to be’ which may serve as an amusement, but is unlikely to cause serious difficulties in a debate. Similarly, the four examples suggested by Walton are largely of deliberately ambiguous phrasings designed to trick hearers into a misinterpretation of offers of a commercial nature. As he notes ‘there definitely is argumentation involved because there is a conflict between two opposed claims’ (2020: 16), yet the examples themselves are not of arguments—they are syntactical ambiguous statements, about the meaning of which we may argue. All of that said, amphiboly cannot be dismissed as an irrelevance. There are genuine examples of mistaken meaning caused by uncertainty of interpretation and they do impact on the flow of discourse. I cite an example at some length below because not only does it provide a good instance of a ‘natural’, unforced amphiboly, but also of how the problem may be resolved and how the author of the sentence reacts to the confusion he has caused. Most interestingly in that last regard, the author in this case reacts both placatingly and aggressively to two different responders, contradicting himself in the process, as is pointed out by the final contributor who arrives to referee. The fact that the user ‘vacantspace’ can claim both that the sentence is clear and admit that it is not in two adjacent posts reveals something of the difficulty of applying argumentation theory to real-life cases. The conversation (complete with spelling errors), in the comments under a Guardian newspaper article (Mahdawi 2019), which touched on issues of personal hygiene, runs thus: 1 The

text of the Amendment reads: ‘A well regulated Militia, being necessary to the security of a free State, the right of the people to keep and bear Arms, shall not be infringed.’ The first noun phrase looks as though it will be a subject, followed by an appositive clause, but it has no verb; the actual subject then appears, strangely separated from its verb by a comma, leaving the Militia clause with no grammatical function in the sentence. To the neutral observer, it seems obvious that the implied subject of the non-finite ‘being’ clause is the militia and that the clause is there because arms are to be borne by militiamen, but it is possible that the ‘right’ phrase could be taken as the implied subject, and there is enough doubt to allow gun enthusiasts to make that different interpretation; hence the debate.

112

8 Fallacies of Language

vacantspace: Personal hygiene and mental health issues are related. Being poor and threadbare was my childhood reality but washing self and clothes was a measure of self respect that goes with desparation and mental decline. Justamentalpatient → vacantspace: Maybe stop washing then. For the sake of your self respect and mental health. Gary Fallon → vacantspace: Surely you meant that “washing self and clothes was a measure of self respect that disappeared with desperation and mental decline.” Or were you just trying to justify the favt you haven’t changed your underwear for six months…… vacantspace → Gary Fallon: Yes, that would have been more appropriate but ‘goes’ and ‘disappears’ are used to replace each other. I used ‘goes’ as in ‘disappears’ not ‘goes’ as in agrees with. Your misreading is understandable as ‘disappears’ is less ambiguous. vacantspace → justamentalpatient: Please read more accurately. Goes, gone, …get it? Goes used in alignment with is an inaccurate modernism. McWeasel → vacantspace: Bit cheeky to tell someone to read more accurately when the meaning they’ve taken from the text is consistent with what you’ve written. Just because it’s not what you meant doesn’t mean it’s not what you wrote—you’ve already acknowledged that goes has two meanings in this context. If you criticised your readers for failing to pick up on contextual cues, that would be more reasonable, there is a generally understood inverse link between depression and self-care so a good faith reading would probably discard the alternative meaning, but my reading of justamentalpatient’s response was that it was a joke about the ambivalence of your first comment rather than a genuine misunderstanding your meaning. Goes used in alignment with is an inaccurate modernism. Should this read ‘Goes used in alignment with with is an inaccuarate modernism.’? If so, that’s not true, ‘goes with’ is a completely standard usage e.g. wine goes with cheese, etc. I don’t intend to provide a full analysis of this excerpt, but a couple of points are worth noting. In the more reasoned reply, vacantspace suggests that the word ‘disappears’ would be ‘less ambiguous’ and McWeasel says that ‘goes has two meanings’, neither realises that the problem is not that goes is ambiguous, but that it can be understood as a single word verb followed by a preposition, as in ‘eyesight goes with age’, or as part of the multi-word verb ‘to go with’ as in the example which McWeasel actually provides, ‘wine goes with cheese’. The uncertainty is actually caused by the possibility of with’s being a stand-alone preposition or part of a prepositional verb, and is, therefore, a structural rather than semantic problem. It is also interesting that vacantspace tries to defend his sentence by suggesting that the alternative reading is ‘an inaccurate modernism’. Quite apart from having just accepted that two readings were possible to another poster, this response commits the two great sins of the grammar pedant: suggesting that being a ‘modernism’ makes

8.1 Aristotle’s List

113

a term unacceptable, and that a grammatical construction can be ‘inaccurate’. Still, however mistaken, this is an appeal to proper usage, illustrating that discussion and negotiation of form of expression, including defending one’s own expression by appeal to a standard, do form a part of real-world debate in our society. Finally, McWeasel shows a good understanding of how textual interpretation works by citing the importance of ‘contextual cues’, but then goes on to misuse the word ‘ambivalence’. For all the expressive and argumentative power of language, a study of actual texts leads to the depressing realisation that even apparently intelligent people are rather poor at using it, and their arguments will generally require a degree of care and attention before they can be properly assessed. These two, homonymy and amphiboly, are both concerned with types of ambiguity, semantic and syntactic respectively, yet they are rather different—homonymy is, as has been pointed out, an essential feature of language, given that one word must be used to signify many things or every single thing would require a unique word to designate it: we could avoid using the same sound for completely different words, but it is equivocation on closely related meanings which is the real problem Amphiboly, on the other hand, might be removed from a language if its grammar were perfectly constructed: having no two cases with the same inflection, not allowing the same form of a word to be used as different parts of speech, for instance. English speakers, I suspect, would rather accept the occasional inconvenience of an amphibolous structure than the introduction of such a system. The third Aristotelean fallacy of language is known as ‘Form of the Expression’. This type of thinking essentially involves making category mistakes about things under the influence of the linguistic forms which refer to them. For example, all verbs behave more or less the same linguistically, yet some refer to states and others to actions; the referents of grammatical subjects are not all things, indeed, they have nothing in common more than that those subjects which refer to them are capable of being predicated. This fallacy has traditionally, including by Aristotle, been categorised as a third type of double-meaning fallacy. Christopher Kirwan disagrees, I think reasonably: ‘figure of speech is not a type of equivocation. The sophist who tries it on is not using one expression with two meanings, and his inexpert respondent does not infer identity of meaning from identity of expression. Rather, he infers identity of meaning-type from identity of expression-type. Form of expression misleads when it does not match form of meaning’ (1979: 42). Whether it should be classed as a type of ambiguity or not, the form of expression fallacy is, I think, one of the greatest importance in the history of ideas. Looked at from the broadest perspective, it includes all of the unfortunate influence which language, and its peculiarities, have on the understanding of reality. Language has developed, in some sense, at least, as an imperfect reflection of the reality of men’s lives, and yet, in many cases, man has been fooled by the nature of language into believing that the nature of reality can be read in that fractured reflection: even that the reflection is a better guide to the nature of the original than the original itself. This is a subject which has exercised many philosophers: Berkeley felt that language was to blame for the very existence of abstract ideas: “had there been no such thing as speech or universal signs, there never had been any thought of abstraction” and believes that a great deal of confusion is caused by the misapprehension that the meanings of words are “precise and settled” (Berkeley 1988: 47). Abstracts, that is

114

8 Fallacies of Language

to say, are nouns in language, with the same range of uses and grammatical features, as concretes, a fact which has led men to believe that they ‘exist’ in the same way. Much more will be said of the influence of linguistic fallacies in general on philosophy, in Chapter 9 below; let it be noted, however, that it is not only philosophers who are influenced in this way: the entire concept of heaven being somewhere above us in the sky, causing believers, and sometimes non-believers alike, to gaze upwards, with a variety of facial expressions, appears to be based on the use of the term ‘heaven’ as a place signifier, with all the grammar that requires: ‘go to heaven’, ‘be in heaven’, and so on, leading to the belief that it must be situated somewhere, like other places we know of. Discovery that the world is round, meaning that we are all looking in different directions, and even the exploration of space, have done little to eliminate this habit: heaven is a place, so it must be somewhere. Krabbe (1998) follows Hamblin in offering an example of fallacy of expression from John Stuart Mill. Mill argues from an analogy with the words visible and audible that somebody’s desiring something makes it desirable, as somebody’s seeing it makes it visible. Yet, they claim, that desirable has a different meaning—the suffix does not refer to ability, but has an evaluative flavour. I disagree, I think this is a simple case of ambiguity—at some level desirable, like all words ending in that suffix—does mean able to be desired. That may not be its primary meaning in modern English, but if Mill has made a mistake, I would suggest it is an ambiguity rather than a category mistake. If he wants to use the conclusion to claim that good (desirable) things can be created because things can be created which are capable of being desired, then he commits an equivocation. A much better example is provided by Krabbe from his own youthful logic: something must exist, he once believed, because if nothing existed, then one thing, i.e. nothing, would exist, therefore something exists. This is a clear case where the grammatical role of the word nothing as a pronoun fulfilling the usual nominal functions of subject, object and complement in sentences has led his younger mind into the trap of thinking that nothing is a thing—despite the quite clear indication in the word that it is not. It is normal to explain to children that nouns refer to things, just as we say that verbs refer to actions; yet in neither case is this a reliable rule. This type of mistake from Mill might also be classed as an example of what Whately called a fallacy of ‘a non-correspondence of paronymous words’ (1857: 118). He gives the examples ‘art and artful, design and designing, faith and faithful’ to which may certainly be added ‘desire and desirable’. This seems a better explanation than the classic fallacy of expression, although it is certainly a species of that type. Schreiber (2003: 47) places Aristotle’s examples of form of expression fallacies into three groups: real category mistakes, where, for example, a substance is confused with a quality; confusion of a particular with a universal in the same category, the basis of the Third Man Argument against Platonic forms; and the confusion of particulars within one category, as in metonymy, where one substance is used in place of another, taken literally. Into this last group he also places confusions caused by the gender of words, where an apparently masculine object is called by a feminine noun. This is not a problem encountered by modern English speakers, but researchers working in other languages might be interested in seeking any examples of the gender of a

8.1 Aristotle’s List

115

language item inducing erroneous thinking, and thus fallacious arguing, about the referent of that item itself. The remaining fallacies listed by Aristotle as being due to language, but not due to double meaning, may present some problems for the modern scholar. Two of them do not appear to be related to language at all, but, as Christopher Tindale notes, ‘Aristotle’s combination of words and division of words differ considerably from their contemporary instantiations as the fallacies of composition and division’ (2007: 58), and, as we saw in the previous chapter, Hurley puts these two into a category of Grammatical Analogy, suggesting he considers them to be rooted in language. In contemporary discussion those two fallacies concern the nature of parts and of wholes: the fallacy of composition occurs when one claims that what is true of one part of something is true of the whole of it, and of division when one claims the reverse, that what is true of a whole is true of each part: a structure built from the strongest concrete may still be weak, a resident of a rich country may still be poor. These fallacies are concerned with substances not the words which describe them and it is important to see that this type of reasoning is not what Aristotle classified as being due to language. Schreiber gives three important starting points in understanding what these fallacies actually are in Aristotle’s writing: ‘(1) They are generated by the arrangement of words in a sentence; (2) they are closely related to fallacies due to Accent; and (3) they must not be confused with amphibolies’ (2003: 58). A brief account of the fallacy of Accent, then, is required. It is a misconception to believe that Aristotle was concerned with mistakes when pronouncing words, although the possibility of pronouncing written symbols in different ways is the root of the problem. Since oral Greek at this period was accented, but written Greek was not, written texts could be misunderstood: this is not a case of ambiguity, since it is not that one word has two meanings, rather, that one set of written letters represents two words. Aristotle’s examples are of disputes over the written text of Homer; he does not suggest that the reasoning of Greeks goes astray because of poorly pronounced language. This is interesting because it actually makes Accent a fallacy of written language and, as such, it is of interest to researchers dealing with unaccented languages: the imperfections and peculiarities of the English spelling system mean that the letters ‘r e a d’ can refer to two different tenses of the same verb, but only when written, there is no ambiguity in the spoken words. To sum up, ‘when differently pronounced words or phrases are transcribable by the identical set of alphabetic symbols, and that transcription plays a role in a fallacious argument, then there is a fallacy due to language but not due to double meaning’ (Schreiber 2003: 59–60) and that is the fallacy of Accent, as is made clear by Aristotle in section 20 of the Sophistical Refutations. The issue with composition and division in language is that certain individual words can be combined, or combinations divided, to alter the meaning of the sentence. To put one of Aristotle’s examples into English: I can write when I’m not writing. This could mean I can both write and not write at the same time, or that I have the capability of writing even when I don’t employ it, with the latter interpretation being the sensible one. The reason this is not an amphiboly is that, Aristotle claims, the oral expression of the sentence would be different in terms of pauses, stress and

116

8 Fallacies of Language

so on, making it not susceptible to a double meaning. Whether the given example works in this way in English or not is unimportant, but it might make for a party game. Crucially, these are also fallacies of writing, though caused by elements of oral language. Punctuation marks are relatively modern inventions, not properly ordered until the efforts of Aldus Manutius the Elder in the early sixteenth century (Truss 2003), and they play an enormous role in the correct division and composition of words in modern texts. Given that ancient scripts were generally written without even putting spaces between the words, the scope for confusion in this way must have been enormous. I would suggest that the current trend in the minimalist application of punctuation marks may lead to a surge in Aristotelian fallacies of composition and division in the near future. There is clearly an analogy to be drawn between equivocation and amphiboly, on the one hand, and accent and composition and division, on the other, with the former in both cases referring to individual words, and the latter longer constructions. The difference between them is that the first two may be found in speech as well as in writing, and the second pair are a product of the transcription of language. It is possible to claim that the distinctions are somewhat fabricated—Kneale and Kneale (1962) suggest that cases such as the ‘can write when not writing’ example are really just equivocations on the meaning of ‘can’—and that there is an element of double meaning about them all, but there is no doubt that they are all cases of uncertainty of meaning. It is also clear that the form of expression fallacy is something different, something deeper and something far more sinister: where uncertainty of meaning can usually be clarified by questioning or context, mistaken conceptions due to form of expression cut deep to the very core of our understanding of the world and how it is related to the language through which we describe it. Finally, it is also worth noting once more that the linguistic phenomena pointed out by Aristotle are not in themselves fallacies, but may be used, deliberately or carelessly, in fallacious argumentation: and this holds true whether fallacies are considered as arguments, which most of his examples are clearly not, or as unacceptable moves within argumentational dialogues, which they do not have to be. What Aristotle has listed, then, should not perhaps be called fallacies, but rather considered as elements of the nature of language, spoken or written, which, when appearing in argumentative discourse, may lead to the drawing of erroneous conclusions. Although modern scholars do not usually begin by dividing poor arguments into those caused by language and those not, linguistic fallacies do continue to receive attention. Some of the many books and papers touching upon the topic are reviewed in the sections below forming the remainder of this chapter. As an introduction to modern thought on the topic I shall make much use of the section on fallacies of language in Christopher Tindale’s book on the appraisal of arguments (2003). This is of particular interest because, as well as a discussion of fallacies of ambiguity, he lists two other poor ‘argument’ types, or rather unacceptable argument moves, in that section which may not universally be considered to fit there: Complex Question and Begging the Question. This leads well into the further consideration of the importance of language in the operation of other traditional fallacies, which follows.

8.2 Ambiguity and Equivocation

117

8.2 Ambiguity and Equivocation An extremely thorough discussion of ambiguity is provided by Jan Albert van Laar in his doctoral dissertation (2003). He carefully defines the phenomenon of ‘active ambiguity’ as opposed to any case of there simply being two possible meanings to a phrase. ‘In order to be actively ambiguous, an expression must admit of several interpretations within its context of use, that is, it must be contextually ambiguous. And in order to be contextually ambiguous it must be semantically ambiguous’ (2003: 35). The emphasis on ‘within’ is van Laar’s and makes it clear that the ambiguity relevant to the analyst is ambiguity in this context, not the general ambiguity of language discussed in earlier chapters. Van Laar refers to semantic ambiguity, a category he takes from van Eemeren and Grootendorst who list it alongside referential ambiguity, where the reference of an expression is insufficiently determined, and syntactic ambiguity, caused by grammatical structure—amphiboly (van Eemeren and Grootendorst 1992a). Active ambiguity, therefore, refers to cases of semantic ambiguity where the multiplicity of meanings is contextually relevant. This is what is generally meant when we talk of ambiguous language being fallacious, but the distinction is rarely clarified. As he is interested in the dialectical effects of ambiguity, van Laar goes on to list four possible events in the ground level discussion, and four corresponding ones in the metadiscussion, caused by instances of active ambiguity. These are: pseudoagreement, concession with divergent readings, pseudodisagreement, and fallacy of equivocation; and, at the meta level, the analysis of each of these (2003: 87). This tabulation is valuable in that it illustrates the ways in which ambiguity can impact on an argument beyond simple equivocations: when one person’s point is not understood properly, it can lead to other difficulties further down the road. For a fallacy of equivocation to be committed, the arguer must actually make use of that ambiguity within the structure of the argument: just using an ambiguous term may make an argument unclear, but it is the action of equivocating, of changing the way the term is to be understood within the course of the argument which is the root of the fallacy. In his Critical Questions concerning potential examples of fallacies of equivocation, Tindale also take into account the effects of the ambiguity. He points out that while we must first establish if there has indeed been a shift of meaning, this alone is not enough for the argument to be written off as fallacious: that shift must affect the quality of the inference, meaning that the conclusion no longer follows from the premises. He suggests two ways in which equivocation within the premises might not be fatal: where there are other premises sufficient to carry the argument, or where the inference is still valid despite the ambiguity. No examples are offered and they are not easy to find or construct, however, this one involves a combination of the two possibilities mentioned: P1: Wild animals, aggressive and out of control, are best avoided. P2: Your dog is wild, aggressive and out of control. Conclusion: Your dog is best avoided.

118

8 Fallacies of Language

The meaning of ‘wild’ here is shifted from ‘undomesticated’ in P1, to ‘crazy’ in P2. This is not catastrophic for the argument though because the salient point is to avoid animals that are aggressive and therefore dangerous, whether they are pets or beasts roaming at large. The argument can be rewritten with the word ‘wild’ omitted and the truth of the premises would be unaffected while the inference remains valid. This is possible since P1 and P2 could both be written out as three separate premises, and the equivocation on ‘wild’ would require the removal of one from each set, leaving sufficient reason to stay away from the dog. Just as an argument’s being fallacious does not mean that its conclusion is false, so an argument which contains an error is not necessarily worthless: as with other types of fault, there are fatal fallacies and non-fatal fallacies.

8.3 The Straw Man as Fallacy of Language Tindale also makes the important point in his discussion of ambiguity that an equivocation has occurred only when the same word is used with different meanings by the same speaker: when two arguers take a word to have different meanings and this is the cause of their disagreement ‘we would judge that they had engaged in a dispute that was really verbal rather than real’ (2003: 60). This refers back to van Laar’s dialectic events, and the consequences which may flow from them. Tindale goes on to suggest that the change in how a word is understood may be part of the construction of a Straw Man: a theme taken up more recently by Visser et al. (2018). They point out, quite reasonably, that ‘misrepresentation is related to rephrasing because a rephrase might constitute means to modifying original speaker’s standpoint in such a way that it is made easier to attack’ (2018: 947). They go on to describe ‘the rephrase relation in terms of the difference between the propositional content of original speaker’s locution and the rephrased statement that has been attacked’ (2018: 948). There are any number of reasons why it may be necessary to restate or rephrase one’s own position or that of an opponent: for clarification, as a reminder, in summary, and so on. What Visser et al. point out is that we do not usually expect any restatement to be made in exactly the same words, although, of course it could be, and this leads to the existence of a continuum ‘from literal repetition of a proposition to expressing an entirely different proposition’ (2018: 949), neither of which can be considered as rephrases, with a variety of more or less accurate expressions of the original content in between, which can be. The authors do not attempt to draw the boundary between the reasonable and the fallacious rephrase; what is clear, however, is that at some point along the continuum the language becomes insufficiently similar to the original for it to be considered a fair representation. This approach shows why this should be considered a linguistic fallacy: while the extreme Straw Man may contain clearly different reasoning to the original argument, the movement along the continuum from copy to free interpretation is dispatched via a gradual shift in the language; a shortening here, a synonym there, until a new proposition has been created. If the borderline could be drawn between the fair and the fallacious, the

8.3 The Straw Man as Fallacy of Language

119

difference would be one word on either side of it. It is, therefore a subtle shift in the language of the restatement that makes a representation of an opponent’s position a misrepresentation.

8.4 Vagueness as a Cause of Fallacy Tindale’s next target is vagueness. This is a very different phenomenon from ambiguity although both result in a lack of certainty about what has actually been stated. As with the earlier linguistic phenomena discussed by Aristotle, vagueness itself is not an argument: it is a quality of language, possibly of arguments. Those who wish to define fallacies as being arguments are not forced to discard the fallacy of vagueness, however. It should simple be named, more precisely, the fallacy of the use of vagueness: that is, an argument which employs vagueness is fallacious, not the vagueness itself. In the linguistic sense, vagueness is an imprecise use of language which renders it impossible to be sure of the speaker’s meaning: an ambiguous sentence may be clear and precise and capable of two interpretations, a vague sentence may be understood in many ways, or not at all. This may be the result of poor language skills, muddled thinking, deliberate obfuscation, or simple under-determination. It is worth remembering, though, that, just like ambiguity, vagueness is an inevitable part of language use: absolute precision would often require tediously long explanations and descriptions, we are generally only as precise as is strictly necessary. Because vagueness is a linguistic phenomenon it requires a thorough linguistic analysis which is well beyond the scope of this chapter.2 Scholars of reasoning have tended to take different kinds of imprecision in language and treated of them together, which is understandable, since their interest is in how that vagueness affects arguments, rather than its semantic or syntactic roots. Johnson & Blair (2006), for example, begin the section of their logic textbook on ‘Fallacies of Language and Meaning’ with two examples of sentences which lack ‘clarity of meaning’: 1. The patient did not fulfil his wellness potential. 2. It was a therapeutic misadventure of the highest degree. (2006: 147) Both of which phrases, they helpfully point out, mean that the patient died. These are cases neither of euphemism, ‘the patient passed away’ is euphemistic yet perfectly clear, nor of ambiguity, there are no two conflicting interpretations; rather the sentences are unclear, and deliberately so. The lack of clarity is not syntactic, both sentences are simple: subject verb object; subject verb complement. Rather it is semantic: the noun phrase ‘wellness potential’ simply isn’t used to mean ‘being alive’, and since ‘wellness’ is in opposition to ‘sickness’, not ‘death’, Johnson and Blair are right to say that the true meaning is being hidden. Similarly, ‘misadventure’ 2 See Sorensen (2001) for a book-length discussion, or his Stanford Encylopedia of Philosophy entry

(2018).

120

8 Fallacies of Language

is not employed to describe unsuccessful treatments, and no normally-competent speaker of English could be expected to understand it as doing so. This is not because it is technical jargon—medical professionals do not usually describe their mistakes as ‘misadventures’ either. On the other hand, Donald Trump’s campaign slogan ‘Make America Great Again’, is short, pithy and, linguistically, couldn’t be clearer; however, the adjective ‘great’ here is vague enough to be susceptible of multiple interpretations—indeed it would be hard for any opponent to disagree with the sentiments it expresses, though his idea of greatness may be very different from President Trump’s. Naturally, this kind of under-determined language is extremely useful for anyone, politician, advertiser, evangelist, who wants to appeal to a large and diverse group of people: such vagueness allows individual interpretations to be projected onto the speaker and also allows that speaker to avoid making more detailed promises, predictions or pronouncements. This does not necessarily have to be a bad thing, though: consider the Nike advertising slogan ‘Just do it’. Whilst the ‘it’ could refer to anything, the context, marketing material from a sports equipment manufacturer, narrows it down to ‘physical activity’, and since the company makes products for use in a wide range of such activities, and since such activity is generally considered a personal and social good, the genius of the phrase is in its very generality. It doesn’t matter here that the ‘it’ is something different for each receiver of the message, the whole point is ‘whatever sport you want to do, don’t think about doing it, just do it!’, meaning that it is relevant for all the market areas in which the company operates. In other fields of communication, of course, such wide generalisation is not acceptable: law-makers, academics, and those drawing up contracts are required to be painstakingly precise: and so it is with argumentation. Sometimes the vagueness may be the result of one word’s having an insufficiently precise meaning. This can be a problem with neologisms where the use of a new term might be simply a way of presenting an old phenomenon as a new one: ‘History provides many examples of persons who, perhaps to emphasize the originality of their views, have introduced new terms where the old ones could have done just as well and perhaps better’ (Naess 1966: 66). New coinages also have a short record of use on which to judge their meaning, and may be used by different groups to mean different things as the process of semantic solidification and ultimately the confirmation of lexicography takes place. A particular problem in argumentation may be caused by the frequently encountered class of terms commonly referred to as ‘fuzzy’ concepts: ‘More often than not, the classes of objects encountered in the real physical world do not have precisely defined criteria of membership’ (Zadeh 1965: 338). Fuzzy concepts have no clear boundaries and lie at the heart of logical problems such as the Sorites paradox. When is a heap no longer a heap? How much hair can a man have and still be bald? The intriguing thing about such words is that they are perfectly intelligible and form a part of successful communications every day: they are not only meant in a fuzzy way when spoken, but they are understood fuzzily too. When I must distinguish between two men and you tell me it’s the bald one I need, I understand you to mean the one with less hair; when you tell me to shift the heap of coal outside in the yard,

8.4 Vagueness as a Cause of Fallacy

121

I don’t come back and say, well, there’s some coal, but I wouldn’t exactly call it a heap myself. These words are not vague in the sense of being difficult to understand, and they are not employed so as to obscure the truth. The potential problems with employing them in arguments, however, are clear. There are different approaches to trying to make sense of the use of apparently fuzzy concepts, and the literature on the topic is wide.3 The view of Sperber and Wilson (1986), defended by Anne Reboul (1989), is that the concepts themselves are not fuzzy, they are only used loosely. From this perspective, ‘bald’ means completely hairless, but it is used loosely to mean anything from that state to a half-covered pate. This is possible because ‘utterances can represent a state of affairs which verifies their forms: they are then descriptive; on the other hand, they can represent other representations […]: they are then interpretive’ (Reboul 1989: 286). So while ‘bald’ has a strict, absolute descriptive meaning, it can be used interpretively in a looser way. This is why we are able to use apparently absolute concepts as though they were scalar: ‘half-dead’ and ‘very much alive’, for instance. Such interpretations, as I mentioned above, are often used for comparative purposes, and there are a whole range of adjectives which do not have any absolute meaning, but work almost exclusively in this way: tall, fat, old, etc. Such concepts may be problematic in argumentation, but they are easily made more precise if necessary, provided the arguer is available to do so. It will also often be the case that a fuzzy word has been used because it is precise enough for current purposes: if I argue that bald men should wear sun screen on their heads on hot days, since sunburn is unpleasant, the salient aspect of their baldness is the presence of some exposed patch of skin, whether there is a ring of hair surrounding it or not is of no consequence to the argument. The Sperber and Wilson solution is less convincing when the paradox is around the use of a noun, since common nouns, at least, don’t tend to have ‘absolute’ meanings. Reboul argues that a word such as ‘heap’ can be considered to have only interpretative meaning; which is a rather unsatisfactory conclusion. If there is a heap and I call it a heap, am I not being descriptive? She suggests that in this case ‘it can presumably be argued that in some cases, at least, the interpretive uses of that word will be very near to literal’ (Reboul 1989: 301); but what that actually means I am not sure. What can be said with certainty is that words such as ‘heap’, ‘pile’, ‘load’, are used quite deliberately to denote an unspecified amount. They cannot easily be precisified: ‘a big pile’, ‘one hell of a heap’, perhaps? They are, therefore, very unlikely to be used in arguments where the exact quantity was a salient feature. Within the rules of Pragma-dialectics, and, one would imagine, any other formulated system of laws for arguing, it is also acceptable for any party to request clarification of points made by any other. An argument made in language which is vague, whether through mangled syntax or under-determined vocabulary, cannot hope to be accepted, but will be sent back to the drawing board, so to speak, by one’s opponents. A second, and more problematic, issue with vagueness, is when it is understood to mean that the argument itself, rather than any particular phrase is unclear: that we 3 For a classic treatment see Lakoff (1973), for more recent developments Belohlavek and Klir (Eds)

(2011).

122

8 Fallacies of Language

cannot be sure what is actually being argued. Tindale notes that the ‘problem here is that we cannot be sure someone intended to argue if we cannot determine what the argument is’ (2003: 65). That is to say, if we cannot identify a clear conclusion stemming from certain premises, then how can we be sure that there is one? How can we know that the passage of speech was intended as an argument when we cannot find the elements that would allow us to recognise it as one? Presumably, we often form a belief from the context of an utterance that it ought to contain an argument, but we cannot quite fathom what that argument is. This is often the case with student essays, for example, where an argument is quite explicitly expected, but can be well hidden. There may be other clues, such as perceptible changes in the prosodic quality of speech,4 or the use of certain vocabulary associated with argumentation, words like ‘because’, ‘therefore’, ‘it follows that’, for instance. All of these concerns lead Tindale to the following definition: ‘vagueness is a problem that impedes us from engaging in arguments’ (2007: 67), which seems a good way of putting it—vagueness is, first and foremost, an obstacle to argument. This perspective would appear to make vagueness more properly belong with the fallacies of process discussed above, despite its clearly being a linguistic phenomenon. At first sight then, we have fallen into the oft-sprung trap of unworkable categorisation: how can fallacies of language be a separate category at the top of the hierarchy if some of them are also part of the bottom category, fallacies of process? This is where it becomes important to see the differences between the types of vagueness which are generally considered together. Since arguments are delivered in language, it is clearly part of the process of arguing to complete that delivery in terms which can be understood by one’s audience, in particular, to make it clear what one is arguing for, and, indeed, that one is actually arguing. It hardly needs to be stated that some linguistic analysis, reading or listening comprehension of the everyday sort, is necessary as a first step before one can assess whether or not due process has been adhered to; however, plenty of arguments exhibiting vagueness will pass this stage of evaluation, because they are still capable of some interpretation by those receiving them. It may be only at a later stage of careful assessment that the possibility of multiple interpretations is discovered, or that a subtler lack of clarity becomes apparent. Examples of this sort will be considered in the final section of the book, where proposals for a full semantic analysis of arguments are set out and tested. As with equivocation, Tindale points out that vagueness is only a problem if it genuinely prevents us from understanding the argument being made. There are many amusing examples of poorly expressed statements where the meaning can be easily guessed. Amazon’s notorious Leadership Principles5 (they have won awards for bad

4 It

is to be hoped that the recent growing interest in Sound Arguments may lead to an analysis of the prosody of spoken argumentation, though as Kjeldsen (2018) notes, studies on the topic so far have kept clear of the empirical. 5 The full horror can be perused at https://www.amazon.jobs/en/principles.

8.4 Vagueness as a Cause of Fallacy

123

English6 ) provide some illustrations. Leaders, they tell us, ‘work to disconfirm their beliefs’ and ‘look around corners for ways to serve customers’, they also ‘do not believe their or their team’s body odor smells of perfume’, the last of which, frankly, is just bizarre; and yet for all their want of elegance and clarity, it isn’t difficult to see what was meant. That being the case, we can still, should we have the courage and the patience, engage in a debate over the true nature of leadership with Amazon executives, who at least admit that their daily use of these principles ‘is just one of the things that makes Amazon peculiar’. At the other extreme are examples where the language used simply has no meaning which can be grasped and communication has broken down altogether.7 It is important to acknowledge here the distinction between clarity and intelligibility. An unintelligible passage of language is not vague—it is unintelligible. Vagueness is not a quality of meaningless language, it is a lack of sufficient clarity in language which can be understood to a degree, only not to the degree of precision necessary for argumentational analysis. Unintelligible statements are filtered out at the stage of process analysis, the really interesting cases are those in which the language, at first sight, appears to be perfectly intelligible and only reveals a hidden uncertainty upon closer examination: these are the types of case with which we shall have to deal in the final sections of this volume.

8.5 Loaded Language Another Aristotelian fallacy which is frequently included in lists of fallacious forms is that of Many Questions, a phenomenon also known as the Loaded, or Complex Question, the classic example of which is ‘Have you stopped beating your wife yet?’. Aristotle placed this amongst the seven sophisms not dependent on language, but Tindale includes discussion of it in the linguistic fallacies section. The obvious problem here is that a question is not itself an argument, and seems hardly likely to appear as either a premise or a conclusion in any written argument. However, as Schreiber notes, ‘questions in dialectic become the premises of arguments’ (2003: 153). At any rate, the principle at work behind such questions is important and is certainly relevant to the discussion of poor or deceptive argumentation. Such a question can be defined thus: ‘it contains an assumption that is hidden but that must be implicitly acknowledged if the respondent is to answer the question. Invariably, such an acknowledgement commits the respondent to a position or claim with which she is at least uncomfortable, and to which she may be adamantly opposed’ (Tindale 2003: 69). It is the notion of the hidden assumption to which we should pay careful attention. 6 See

https://idler.co.uk/article/news-harry-styles-wins-our-first-ever-good-grammar-award/ for details. 7 See the examples from writings in Continental Philosophy highlighted in Smith (2006), or the Zen stories compiled in Reps (1971) where language is used only to show the limitations of linguistic meaning.

124

8 Fallacies of Language

Take a simple example of an argument pattern which may be called argument from promise or expectation: P1: As an American, you want America to be great. P2: Donald Trump promises changes to make America great again. Conclusion: You should vote for Donald Trump. Whatever else may be wrong with this argument, it slips in the hidden assumption contained in Trump’s loaded slogan through the use of the word ‘again’: that America was great in the past, but this is no longer the case; which in itself implies that changes are needed in America to restore its greatness. A voter may well agree with premise 1, and acknowledge the truth of premise 2, and, plausibly, believe Trump is sincere and able to do what he says, and still not accept the argument because he does not allow the hidden assumption. On the contrary, he believes that America is great, and no changes are necessary to secure that status. Such cases are related to the concepts of argumentativity of language, and implicature, discussed in Chapter 3. So while the fallacy of complex questions may not actually refer to fallacious arguments, it highlights an important element of the semantic analysis towards which this study is ultimately leading. Johnson and Blair discuss ‘loaded’ language from a different angle. Rather than requiring acknowledgement of an implied statement, what they name ‘freeloading terms’ assume acceptance of an attitude: ‘words and terms are loaded in the sense that many of their uses – not all – carry with them an evaluative component. Many words are by convention loaded with emotional and psychological freight’ (2006: 160). Without wanting to get into a debate on the nature of emotions and how we can evaluate them, it is worth looking at the role which evaluative and emotional language plays within the structure of arguments. Michael Gilbert has long argued for greater attention to be paid to emotional arguments by the argumentation theory community. He suggests that ‘the utilization of emotion in arguments in part or in whole, is perfectly rational’ (Gilbert 2004: 246) and bemoans the attitude of philosophers in general which he characterises as: ‘Persuasion and its handmaiden–emotion–are bad, and convincing with its valet, reason, are good’ a position he rejects, since ‘emotional messages do carry information rather than merely colour the dialectical information contained therein, and […] emotional information is used whenever we are convincing as well as persuading’ (Gilbert 2004: 261).8 In a similar vein, in his 2014 paper on value-based reasoning, Fabrizio Macagno describes the dual nature of emotive language used in an argumentative setting. When analysing instances of such vocabulary, he suggests that: ‘it is necessary to take into consideration their two distinct and connected dimensions: their logical functions as 8 Naess

makes his opinion clear: ‘Speech designed to arouse men to action or to have a direct influence on their behaviour differs in kind from speech used purely in the exchange of ideas. To combine the two functions defeats the aim of each, namely, commitment on the one hand and clarity of thought on the other’ (1966: 121).

8.5 Loaded Language

125

implicit and condensed arguments, and their rhetorical effect, consisting in arousing emotions’ (Macagno 2014: 107). It is the idea of the condensed argument—‘emotive words can be short-cuts, simplifications of chains of arguments’ (Macagno 2014: 118)—which is of most interest. If we describe the killing of a man as a ‘murder’ we are thereby arguing for the censure of the act and the punishment of the perpetrator. There are two obvious advantages for public speakers in employing such short-cuts: one is that they save time, a very important consideration in certain time-constrained situations, and the second is that they are more direct and considerably easier to follow than a fully drawn out argument delivered in carefully neutral language. The audience already knows what it thinks about justice, corruption, and dying children; on such issues we can assume a number of shared value-based premises across the majority of those listening which can justify the statements made using those phrases. The degree to which such practices are acceptable at higher levels of debate, where time constraints are not a factor and assumptions about audience beliefs are harder to make, is debatable. Still, as Macagno concludes when stating that emotional arguments can be assessed just like any others, provided that they are allowed to rest on a certain kind of evaluative premise: ‘Emotional reasoning is still a form of reasoning which can be assessed as grounded or not on shared premises […] Sometimes emotive words are needed in order to summarize in a condensed argument a more complex type of reasoning’ (Macagno 2014: 118). These comments make it clear that although the use of emotionally or evaluatively loaded language is generally considered to be poor practice in argumentation theory, vocabulary which is marked in this way can also be used to convey serious argumentative content. Such language, however, needs to be treated with great caution—it too has inherent argumentativity, since it argues for the assumptions necessary for its acceptance. The slogan sometimes employed in the cause of vegetarianism, ‘Meat is Murder’ argues, through its use of the evaluative term ‘murder’ that the whole process of butchering animals for meat is unethical and unjustified. The word itself requires this premise; this is, interestingly, a kind of backwards argumentation from a conclusion to the premise which is implied by it. Therefore, while the slogan uses evaluative language to greatly condense its argument, which is useful, it is still fallacious to the extent that it demands a commitment to a controversial premise. This particular example is also guilty of begging the question when it is employed in an argument against meat-eating on ethical grounds, which leads us in to the next section.

8.6 Begging the Question Begging the question is also listed by Tindale as a fallacy of language. The example from the previous paragraph shows why this is a reasonable treatment: the question in the meat-eating debate is whether or not raising and killing animals for food is

126

8 Fallacies of Language

morally acceptable. The statement ‘Meat is Murder’ requires, as noted above, the premise that such killing is wrong. The argument might look like this: P1: Slaughtering animals for their meat is unethical and unjustified. P2: Unethical and unjustified killing is murder. Conclusion: (The slaughtering of animals for their) meat is murder. And yet, since the entire debate revolves around the truth of the first premise, this argument simply begs the question. Indeed, since we know the definition of ‘murder’, the second premise, the conclusion and the first premise, in fact, say exactly the same thing. Aristotle discussed the fallacy of begging the question, or petitio principii, but placed it in the category of fallacies independent of language. Since it involves confusion over the separation of premises and conclusions, it could be thought of as a mistake with the reasoning part of the argument, rather than its expression, but examples such as the one above show how the load of certain linguistic items can be the driver of this fallacy. It is also true, of course, that circular reasoning of this type is not logically invalid: p therefore p, is perfectly acceptable to the formal logician. John Woods and Douglas Walton, in one of their early papers, note that ‘petitio, like other informal fallacies, occupies something of a logical twilight zone’ (1975: 107). They go on to make a distinction between two ways in which questions may be begged, a division which is followed by Tindale. There is, they state, an Equivalency Conception and a Dependency Conception of the fallacy. In the former, ‘the conclusion is equivalent, or even identical to one of the premisses’ and the in the latter ‘the conclusion is presupposed by a premiss or where some premiss actually rests on the conclusion’ (1975: 108). This can be paraphrased as a conclusion is either equivalent to a premise, or entails one; and the authors acknowledge that the distinction between the two is not entirely obvious and may be illusory. This seems correct, but is deeply problematic. Isn’t it the case that in any deductively valid argument, the premises and conclusion are equivalent? Isn’t the sentence ‘the radio is either on or off, and it’s not on’ somehow the same as ‘the radio is off’? This is a simple disjunctive syllogism: should it be considered circular reasoning? Deductive reasoning cannot create anything new, the conclusion must be in some way the equivalent of the premises. Woods and Walton note that: ‘For a logically omniscient being (Hintikka’s Superknower) all arguments would be circular in which the premiss is equivalent to the conclusion. For the rest of us, whether or not such arguments are circular depends on how transparent the equivalences are’ (1975: 117). This degree of transparency will vary from person to person, appearing to make the fallacy one of relative perception. Arguing that John is unmarried, therefore John is a bachelor, is only circular if I know what the word bachelor means. There is, I think, a different way of framing things so as to preserve the notion that such arguments are special and are fallacious in a more objective way, and that is by returning to the notion of the question which is being begged. I wrote above that the vegetarian argument concluding that meat is murder was circular when used in a debate about the morality of eating meat, but it is not circular, nor fallacious at all, for those who share the evaluative background of the speaker in the way discussed

8.6 Begging the Question

127

by Macagno. The fallaciousness is not purely subjective, however, because what the argument does wrong is to use the answer to the question as a premise, and then again as a conclusion, and this is the fundamental difference between the formal logical, and the real-world, informal conception of argument. A circular argument is one which uses an answer to the question being addressed in its premises: the arguer knows that this premise is what is under debate, but assumes a resolution of that debate within his own attempt to provide that resolution. If I say as above, ‘the radio is either on or off, and it’s not on, therefore it’s off’ to someone with no opinion about the radio one way or the other, no question is being begged; but, if my friend has just said to me ‘the radio is on’, then, using the premise that it is not on in order to prove that it is off, is begging the question at hand, which is—is the radio on or not? Simply put, in order for a question to be begged, there must be a question: and relevant argumentation must address that question and not assume an answer to it. As Kienpointner notices, one particularly interesting variety of linguistically induced question begging comes from the use of tautologies (1996: 490). For example, my Polish neighbours can explain any bad behaviour of their dogs by stating ‘pies to pies’—‘a dog is a dog’—apparently claiming that the behaviour is normal in a dog because it is normal in a dog, when the very question I have raised is whether this behaviour is part of the normally accepted behaviour of dogs. Of course, it would be ridiculous to expect a dog to be a man, or even a cat, but must I accept my neighbour’s understanding of what it is to be a dog? Dogs may be a relatively unimportant case, but one of Kienpointner’s examples is ‘War is War’; in so-saying one is claiming that the current situation is war, and that all wars are the same, when one’s interlocutor may well have been trying to clarify the situation and introduce a more nuanced understanding of current events, raising the issue of whether this situation is, in fact, similar to other conflicts. Kienpointner cites Okamoto who notes that by stating an equative tautology—A is A—‘all argument over categorization is refused, and category immutability is simply insisted on with no logical explanation’ (Okamoto 1993: 439). The degree to which begging the question can be considered a linguistic fallacy and not a procedural fallacy will, as with questions of vagueness, depend upon the subtlety of the case. Johnson and Blair do not include begging the question in their chapter on fallacies of language, nor do they state explicitly that it should be so considered, but the explanation of the examples they give, and why they consider them to be fallacious, makes it very clear: when considering a premise and conclusion pair they note that ‘these two phrases express the very same notion in different words […] these statements are merely semantic variations on the same theme’ (2006: 82): and, tellingly, a few lines earlier they point out that: ‘The flowery phrasing of the argument serves to disguise that it begs the question’. In situations where a premise is restated exactly as premise and conclusion, where an answer to the question at hand is clearly assumed in giving an answer to that question, then the fallacy is obvious and it is a fallacy of process that has been committed, since the argument itself is irrelevant in the resolution of the question. More interesting cases are those, analogously to instances of lack of clarity, where the premise and the conclusion are not stated in the same words and the connection between them must be teased out through a

128

8 Fallacies of Language

careful analysis of what may sometimes be ‘flowery language’. That is to say, when the equivalence or dependency of the two elements is somewhat hidden within the argumentative load of the words or phrases used. Such cases clearly deserve to be considered as falling within the linguistic sphere, and will only be uncovered by an analysis of that sort.

8.7 Hedging Gary Jason (1988) is also keen to expand the range of linguistic fallacies, and suggests that too much emphasis in this area has been put upon ambiguity as a source of error. As well as mentioning that both Straw Man and Begging the Question may be based in a misuse of language, he proposes Hedging as a fallacy type worthy of individual recognition. The phenomenon is described in two ways. In his first, made-up, examples, he shows how the proposer of a standpoint may gradually back away from what was originally claimed by hedging the statement when confronted with counter-evidence, so that the claim ‘Henry is a totally selfish man’ is eventually weakened to: ‘Well, about matters that affect his wallet he can be very selfish.’ The fallacious move here is described thus: ‘a claim which is reasonably qualified at the beginning of the dialogue gets systematically weakened as the dialogue proceeds to avoid the thrust of counterevidence’ (Jason 1988: 169). I am not at all sure that there is anything fallacious about adding qualification to a claim: the speaker does not try to redefine what selfish means, he simply narrows the scope of his standpoint. If the counter-evidence continues to mount, he may either narrow it into triviality or withdraw it. Jason seems to suggest that the proper behaviour would be an immediate withdrawal, but a certain amount of hedging is surely a reasonable reaction to a certain degree of counter-evidence. Imagine this dialogue: A: You’re always late! B: I was on time last Friday. A: Well, you’re usually late. Get a watch! Would the following be a more reasonable response? A: Oh, yes, you were. I retract my statement and apologise. If there is a fallacy of hedging, then, it is one of over-hedging, and it may well be better considered as one of process, since there is nothing wrong with modifying language; the ‘crime’ here is of failing to retract a standpoint which has been disproved. The further examples he provides from actual political debate are of presidential candidates talking round a subject and deflecting from the point with irrelevant information, and don’t seem to contain the same type of moves at all. The second variety of hedging is exemplified by product guarantees and involves the qualification of a claim to the point at which it becomes meaningless, i.e. so many terms and conditions are attached to a guarantee that it becomes no guarantee at all. Again, although this practice may be somewhat dishonest, especially if the

8.7 Hedging

129

qualifications are well-hidden from the customer, it is hard to see why it is actually fallacious, especially as guarantees are not normally considered as arguments. Interestingly enough, Jason appears to have got onto the wrong side of an amphiboly in his discussion of this: he cites a guarantee with the phrase ‘will be replaced with like fresh film or purchase price refunded, at our option’ which he describes as being ‘hedged to be optional at the will of the manufacturer’, in contrast to a second clearer one which states that the manufacturer ‘will replace free of charge or refund full purchase price’. In the first example, the ‘option’ could be understood as applying to the whole process of compensating the customer, but this seems unlikely, rather it refers to whether they are given a replacement or a refund, where the second example leaves unstated who gets to choose. The second example, although written in clearer language, is, therefore, guilty of a certain vagueness. The comparison is a little unfair in any case, since the first is a careful text in legal English and the second a marketing claim, which is almost certainly qualified elsewhere. There are two interesting points which represent afterthoughts in the short paper, but might well have been expanded to good effect. The first is Jason’s mention of the fallacy of reification, or misplaced concreteness, which will be discussed at length in the next chapter as a common philosophical fallacy connected with abstract concepts. The second appears in a final endnote, where he writes: ‘One might even take the extreme position that there are no separate “fallacies of language” per se but rather that fallacious reasoning can occur because of misuse of language’ (1988: 175). Provided one takes a sufficiently wide understanding of ‘reasoning’ here, I do not think that this is an extreme position at all. That understanding would need to include at least some of the elements of what I refer to as process: vague language, for example, prevents the process from progressing; misrepresentation of opponents makes one’s arguments irrelevant. Jason considers ambiguity the hardest fallacy to account for, but it seems to me the easiest: in equivocations, use is made of a common term, which is not in fact a common term, the simple deductive argument ‘a is b, and b is c, so a is c’, becomes the illogical ‘a is b, and c is d, so a is d’. Other fallacies which are discussed in this chapter and the next can generally be seen as cases of the arguer not arguing for what he thinks he is arguing for, or using language which is essentially meaningless, which renders his reasoning meaningless too. I do not want to pursue this further here for two reasons. One is that the definition of ‘reasoning’ is a question already discussed and, for our purposes, decided upon. As I have said, to define all language fallacies as really fallacies of reasoning would require a broader interpretation than I have proposed. The second is that I am interested here in the stages of argumentation: reasoning, expression, and process, and how fallacies can come to light when examining any of the three. I am at pains to express throughout the interconnections amongst them, and leave no doubt as to the fact that errors at one stage may at times only be revealed through the analysis of later stages. Whether all the errors revealed by linguistic analysis can be related back to errors in process and reasoning, is, for my purposes here, a little beside the point. There are three more linguistic phenomena which affect the acceptability of arguments, though they are not usually found listed as fallacies of language, which I need to discuss here as they will form a part of the final analysis which is employed in the

130

8 Fallacies of Language

last sections of this work: these are Persuasive Definition, Self-defeating Statement, and Semantic Incompatibility.

8.8 Persuasive Definition Persuasive Definition is a name introduced by Charles Stevenson, and he described the aim of those who employ such definition as being to ‘alter the descriptive meaning of a term’ whilst maintaining unchanged ‘the term’s emotive meaning’ (1944: 210). The adjectives descriptive and emotive are not particularly helpful, and in many ways not very important, here, as Andrew Aberdein has noted (2006). What is important is the change in the definition of a word whilst still maintaining that it is, essentially, the same word with the same range of uses. An Englishman returning from the Continent may sit down to his afternoon tea and exclaim his relief at returning to ‘civilisation’, but he is, by so doing, defining ‘civilisation’ rather differently to the writer of his dictionary. This can be amusing when done with self-awareness; it is merely a form of exaggeration, but that is not always the case. Aberdein lists some of the varieties of fallacies of this sort: most interestingly the ‘No-true-Scotsman’ and the more general practice of ‘Dissociation’. The ‘No-true-Scotsman’ fallacy derives from the work of Antony Flew (1975) and the example of a Scotsman taking sugar in his porridge to disprove the dictum that no Scotsman does so. The saying can be salvaged by redefining the terms and stating that no true Scotsman does so. The definition of Scotsman is altered to mean not just a member of a nation, but a member of that nation who exhibits a certain characteristic, to wit, not putting sugar in his porridge. Thus-defined, the saying becomes self-evidently true. There is something akin to begging the question here— accepting that no true Scotsman puts sugar in his porridge entails accepting the premise that any Scot who does so is not really a Scot. There is also the likelihood of an equivocation if the revised statement is used as a premise itself: P1: No true Scotsman puts sugar on his porridge. P2: Jimmy is a Scotsman. Conclusion: Jimmy doesn’t put sugar on his porridge. When the re-defined Scotsman and the normal Scotsman are both used, then an equivocation has been committed, even though the switch in meaning has been flagged with the qualification ‘true’. A very similar move is labelled in the posthumously published work of Imre Lakatos (1976) as ‘monster barring’. On Lakatos’s description this is an ad hoc change in a definition of terms in order to side-step a possible counter-example to one’s theory. Aberdein calls this move ‘gerrymandering a term’, a description which can also be applied to the wider practice of Dissociation discussed by Perelman and Olbrechts-Tyteca (1969). Dissociation means the division of a concept into two: in the ‘No True Scotsman’ move, this division is between the ‘true’ meaning and some other, presumably ‘false’ or merely ‘apparent’ one. Other divisions are possible, of

8.8 Persuasive Definition

131

course, between a ‘general’ and a ‘technical’ meaning of a word, for instance, and where this is done explicitly and with justification there is no need to condemn it, but even in the best intentioned instances of dissociation, and any redefinition which contains a normative element, the ‘true’ ‘real’ or even ‘recommended’ meaning of a word, is likely to be unacceptable to at least some of those whose preferred definition is being usurped. Certainly, anyone employing such redefinitions in earnest has misunderstood the fundamental nature of language: however one conceives of the meanings of words, one cannot simply change them at will. Persuasive definitions, then, can lead to argument failures in two ways: they may simply be rejected by the audience or they may lead to equivocation. An example of how such differences in definition can be important in public life can be found with the word ‘poverty’. My rather old-fashioned Shorter Oxford English Dictionary (1983) has as its primary definition ‘The condition of having little or no wealth or material possessions; indigence, destitution, want’. The United Kingdom government’s Office for National Statistics, on the other hand, suggests: individuals are considered to be experiencing relative poverty if they live in a household with an equivalised disposable income that falls below 60% of the national median in the current year. This is a relative low income measure also referred to as the “at risk of poverty” rate, which measures income compared with other people. As such, being at risk of poverty does not necessarily imply a low standard of living. (ONS 2017, Sect. 2)

This definition is at least honest, the last line makes it clear that ‘relative poverty’ doesn’t actually mean ‘poverty’ at all, it is simply an invented measure of inequality. Whilst an exact technical definition is obviously required for the task of measurement and comparison, a measure of poverty which does not imply a low standard of living, is a clear case of ‘gerrymandering’ of meaning. The ONS report itself simply shows figures for various countries using the metric it describes; however, when journalists and politicians use the statistics to make claims and accusations they knowingly use the word ‘poverty’ in this technical sense, whilst being aware that it will be received by many in their audience in its normal sense of destitution and want. This is a kind of implied equivocation: the person making the argument that government policy should be changed in order to reduce poverty may be consistently using poverty in its technical meaning: nonetheless, the argument is deceitful since the speaker hopes and expects that the person hearing it will understand something different by the word. A report from the Rowntree Foundation using a similar measure to that of the ONS was reported in the Guardian newspaper with the following: ‘Margaret Greenwood, the shadow work and pensions secretary, said the report should be a wake-up call for the government. “There is something seriously wrong when the number of people in work in poverty is increasing faster than employment,” she said’ (Partington 2018). This statement clearly implies that the reduction of poverty is a responsibility of the government, and the opposition spokeswoman quoted may believe that that includes poverty defined as inequality, but she is certainly aware that many voters might not agree with that and that most voters hearing her words won’t interpret poverty as per the technical measure. This particular example could also be defined as a fallacy of loaded language, since ‘poverty’ is actually being used as a more emotive substitute for ‘inequality’.

132

8 Fallacies of Language

8.9 Self-Defeating Statement Another interesting phenomenon of language which may lead to fallacious forms treading the line between logical and purely linguistic error has been named, very infelicitously I think, the ‘Lord Scroop fallacy’ by Herman Stark (2000). The name is a somewhat obscure reference to the traitorous nobleman from Shakespeare’s Henry V; the point being that statements which commit this fallacy turn in upon themselves and contain the cause of their own destruction within them. For that reason, I refer to such cases as Self-defeating Statements. Stark makes it clear that he is concerned about the use of these seductive, but ultimately empty, phrases by his students, and see them as undermining the intellectual effort required to make true academic progress. He gives examples such as ‘Everything is a matter of opinion’ and ‘There is no truth’, which he describes as ‘dogmas’. These dogmas are easily disproved—either it is true that there is no truth, in which case the statement is disproved, or it is just false. There is no suggestion anywhere in the paper that this should be considered a linguistic fallacy. Indeed, Stark makes it clear on several occasions that he considers it a logical error, although admitting that ‘Scroop does not involve a mistake in reasoning’ (2000: 255), and likens it to Begging the Question, since both constitute ‘a failure in argumentation, and this failure also cannot be accounted for in terms of an invalid inference’ (2000: 256). If one were to claim that any fallacy brought about by the meanings of words should be considered a fallacy of language, then, in natural argumentation at least, there would be no other sort. However, since the conclusions drawn from these self-defeating premises are not invalid, and the statements themselves are false, not because they refer to an incorrect understanding or incomplete knowledge of the facts, but because their semantics are somehow incompatible with themselves, anyone making such statements clearly, at a deeper level, does not understand what they mean: the falsity of the statement is, to a degree, hidden in the language, and, again, it is only a careful analysis of the language of an argument which can expose this. Some, indeed, are idioms in their own right and commonly used in everyday argument: ‘There are two sides to every story’, for example. The wisdom contained in the figurative language may be lost when it is taken literally: when a person bears witness against another, there is always a second side to the ‘story’; when photographs from space show that the Earth is not flat, but, indeed, spherical, there is not. I am not interested here, or indeed anywhere in this chapter, in arguing that the phenomena discussed, the fallacies or paths to fallacious inference, should be categorised as linguistic fallacies. As I stated when defining argumentation, this is one way of conceptualising arguments and my interest is mainly in the process of evaluation. Some self-defeating premises may be obvious enough to lead to the rejection of the argument at the process stage; still others may be found out at the reasoning stage; and yet, some, I believe, are so cleverly, so intricately, bound and folded within the language in which they are expressed, that their existence will only be brought to light at the final stage of analysis, where the linguistic fine print, as it were, is examined.

8.10 Semantic Incompatibility

133

8.10 Semantic Incompatibility The final fallacy of language which I shall discuss is one which I intend to introduce, not having seen it described elsewhere. I shall consider as fallacious any argument which rests on the employment of a Semantic Incompatibility in its premises or discovers one in its conclusion. Instances of the former are less interesting as they are no different from other forms of vagueness or, at worst, meaninglessness; occurrences of the latter, however, raise fascinating questions about the fundamental relationship between language and ‘logical’ argument. A semantic incompatibility I define as the co-dependence in a sentence of two linguistic items, the meanings of which disallow their collocation. This might be easy and obvious such as ‘He was a married bachelor’, or subtler, ‘I choose not to believe in free will’. In the second case, although both points are contestable, we may find two incompatibilities: one cannot choose one’s beliefs, and one cannot choose to deny that one has any choice! Certainly, such phrases can be used for comic effect, but an argument which relies on one, or produces one as its conclusion cannot be accepted. A fine example of this, which was of great topical interest in the United Kingdom as it exited the European Union, was the frequently repeated claim that leaving with ‘no deal’ was not an option. This claim was made in the Labour Party manifesto for the 2017 General Election and repeated by a great many politicians in the months leading up to the withdrawal date. In the section of the manifesto entitled “Brexit”, the following statement appeared: “Labour recognises that leaving the EU with ‘no deal’ is the worst possible deal for Britain and that it would do damage to our economy and trade. We will reject ‘no deal’ as a viable option” (Labour Party 2017: 24). Describing ‘no deal’ as ‘the worst possible deal’ is a little odd, but more important is the combination of the words ‘reject’ and ‘no deal’. These items are simply incompatible—a lack of a deal is not the kind of thing which can be rejected. When that manifesto was written, no deal was in existence; once one had been concluded between the EU and the British government, it was roundly rejected by the same people who confidently rejected ‘no deal’. The argument can be set out thus: P1:

Reaching no deal with the EU on trade would be against the interests of the UK. P2: We must reject what is against the interests of the UK. Conclusion: We must reject ‘no deal’. What makes this so worthy of consideration is the fact that the premises look reasonable—only a small group of British politicians rejected premise 1—and the inference from premises to conclusion also looks quite valid. Yet the conclusion is nonsense: even if it were interpreted as a readiness to accept absolutely any deal whatsoever, and clearly that is not what was meant as the subsequently negotiated Withdrawal Agreement was not accepted by the Labour Party, one still cannot refuse to accept that one’s partner in establishing the deal ultimately declines to offer one. The persistent admirer’s ‘I won’t take no for an answer’ is not a serious policy position. Any argument which generates this kind of inconsistency must be rejected, simply because it cannot be accepted. This fallacy may be treated as a special case of

134

8 Fallacies of Language

vagueness—the conclusion seems to make sense, but it simply isn’t clear what it actually means. Supporters of the position advocated by this argument, or any other which commits the same fallacy, may rush to complain that it can be salvaged through rephrasing, and the same can be said for many of the arguments guilty of the faults discussed in this chapter. Which is perfectly fine: if your argument has fallen into linguistic fallacy and you can save it through re-wording, then that is exactly what you should do, and when it has been properly reconstructed it may be considered again. The mere claim that it could be saved in this way is not, however, sufficient; I cannot see, for instance, how the ‘no deal’ rejection argument can be rephrased successfully without being seriously watered-down. Ultimately, as discussed in Chapter 6, the defence that an objection to one’s argument is ‘just semantics’ is no defence at all—would anyone say ‘it’s only meaning’?

Chapter 9

Linguistic Fallacies in Philosophy

In the preceding chapters, the assertion has been made more than once that fallacies of language are a particular vulnerability at the highest levels of argumentational discourse. Even those who are determined to be rational and have a strong understanding of the laws of inference; those who have their emotions in check, and their prejudices well-disguised; those who would seek out the truth from no other motive than a love of knowledge and wisdom; even they must put forward their ideas in language and are, therefore, in danger of falling into the many traps that language has set. Philosophy, the archetypal practice of disinterested reasoning, abounds in accusations of misunderstanding and confusion over the meaning of words and their use. So much so that one might be tempted to write off the whole field as nothing more than a series of arguments about language conducted by people who don’t understand how it works. Whether or not that is a fair assessment is a question for the concluding section of this chapter. Before then, I shall consider the concerns of Wittgenstein and Berkeley, G.E. Moores’s concept of the Naturalistic Fallacy, and, firstly, Leonard Nelson’s theory of the Philosophical Fallacy. I shall then discuss certain other philosophical positions in order to show where issues with language have led their advocates into error. It should be remembered, of course, that there will be no attempt made to fully discuss the philosophical questions that these arguments address, and for this reason there is no review of current literature on their various topics. We are concerned with argumentation and the point of interest here is in the ways in which philosophers have made errors, the ways in which other philosophers have criticised their reasoning, and, especially, the fact that these criticisms can all be reduced to special instances of the fallacies of language described in the previous chapter. This is true even when those who have highlighted such errors have done so without making the linguistic element explicit, and, often, while their real motive was to move from the apparent error of their rivals to the, subsequently, obvious truth of their own position.

© Springer Nature Switzerland AG 2021 M. Hinton, Evaluating the Language of Argument, Argumentation Library 37, https://doi.org/10.1007/978-3-030-61694-6_9

135

136

9 Linguistic Fallacies in Philosophy

9.1 Language in Philosophy The complaints raised in the first part of the chapter, by philosophers against their fellows, follow the tradition of Aristotle’s Sophistical refutations, but are treated here separately from other discussions of fallacy because these are claimed to be made specifically whilst practising philosophy, rather than in more general reasoning. The examples of argument which are critically examined both by the philosophers mentioned, and later in my own analysis, are largely in support of well-known, historical theories. These are the theories which, in modified form, are the basis of more modern philosophy anyway, and are also easier to introduce without long explanation as they will be familiar to many readers. I take it as uncontroversial that philosophy may be considered the highest form of argument in terms of its precision, its sophistication, and the scrutiny to which both its reasoning and expression are subjected. Philosophy is also the deepest form of argument in that it examines the fundamental beliefs upon which all other reasoning rests: political and legal debate may appear to have more immediate impacts on the lives of the great mass of people than obscure philosophical theorising, but the authority of politicians and judges is granted on the basis of the justifications of rights, duties, and ethical standards which are made by philosophers, not lawyers or parliamentarians. I do not assume, as will become obvious, that philosophers are inherently better at arguing than others, merely that the procedure of philosophical argumentation is the most rarefied and refined, and the most likely to turn on points of language. It is at once the most profound and the least widespread; its concerns are the most universal, its practice the most restricted. This special status leads to two particular dangers: one is a pretentious obscurantism, which pleads that the nature of its topics is such that straightforward language simply will not suffice; and the other is a certain insularity which feeds on jargon and self-reference, and may result in the creation of a language within a language. When this happens, philosophers may discover that they are talking only to other philosophers and only about problems of philosophy, rather than using the tools of philosophy to discuss the problems of life and the universe. I take this to be an undesirable state of affairs, just as I take it that all that is truly wise is susceptible of clear and simple expression: some philosophy may be difficult, but we must be always inquisitive as to whether that is a result of the great complexity and profundity of its content, or a mere product of its being poorly expressed. One more general comment that is worth making concerns the attitude to language which comes through in the following paragraphs. The philosopher who would be more than a mystical hermit and see his ideas spread and discussed, has but one choice of medium: he must take his reasoning and turn it into arguments by the employment of language. Yet, deep thinkers have long had a difficult relationship with words: there is a belief in their power, but still a strong suspicion of their inadequacies. If only language could be made to behave correctly, if only words were properly understood, then the truth would become clear. The opening verse of the classical Taoist scripture, the Tao Te Ching (1963), runs thus:

9.1 Language in Philosophy

137

The way that can be spoken of Is not the constant way; The way that can be named Is not the constant name. The author, Lao Tzu, who more than likely never actually existed, clearly understood the folly of committing his wisdom to language, he recognised that language was not up to the task of describing that which he wished to describe, the Tao, the constant way, and yet, he, or whoever it was who actually contributed to the collection of passages which make up the work, proceeded to add a further 80 verses. These verses all speak of the way that can’t be spoken of, as if no-one had ever suggested that that might be a difficult thing to do. The Tao itself is a special case: it isn’t only, in Lockean terms, an abstract idea formed from a complex set of qualities, it is an abstract formed from a number of abstracts. As such, Lao Tzu recognises that while we may give it a name, the use of that name does not equate to an understanding of a parallel concept. He is not trying to redefine a term already used by men, he is trying to introduce an idea which men cannot cleanly grasp in their normal thoughts and for which, therefore, they have no words. In this way, the Tao Te Ching avoids the philosophical fallacies of language described immediately below, because it does not attempt to take common terms of language and tell us that they really mean something else; and yet, paradoxically, because it tries to define a term which is not commonly understood it slips into what can only be called mysticism: ‘Turning back is how the way moves;/Weakness is the means the way employs’ (Lao Tzu 1963: 101, verse XL); which is hardly the sort of thing one could sensibly argue about.

9.2 Philosophical Fallacies of Definition After these general remarks, I turn now to the critical analysis which some philosophers have offered of the work of others and of the practice of philosophising. There is a long tradition: John Locke noted wryly that: Though the Schools, and Men of Argument would perhaps take it amiss to have any thing offered, to abate the length or lessen the number of their disputes; yet, methinks those who pretend seriously to search after, or maintain Truth, should think themselves obliged to study, how they might deliver themselves without Obscurity, Doubtfulness, or Equivocation. (Locke 1690, III, XI, 3 [1975: 509])

The recently translated work of Leonard Nelson (2016) is a good place to start, since in the series of lectures of which it consists, he sets out to build an entire theory of philosophical fallacies. Not only that, but he does so from the viewpoint that reasoned argumentation is of paramount importance in the pursuit of truth, which should be enough to make him a hero of argumentation theorists everywhere. Nelson begins, in the first lecture, by attacking what he sees as a rise in ‘artistic’ philosophy, by which he means philosophy not based on the hard work of careful reasoning, but on the easy course of intuition and following the ‘spirit of the times’. He gives an example

138

9 Linguistic Fallacies in Philosophy

the work of Oswald Spengler ([1918] 1991) whose ‘wordless understanding’ finds its expression in ‘over 600 pages of wordy text’ (2016: 25). He refers twice to the length of Spengler’s book, making it clear that Nelson himself is suspicious of such a proliferation of words. That more words usually means less clarity is one of the many ironies of language we have had cause to note along the course of this volume. For Nelson, truth must be reached through reasoning, and progress comes through what he calls the ‘dialectical field’, by which he means philosophical argumentation, and specifically through the application of logic. This leads him to use the phrase ‘dialectical error’ interchangeably with ‘fallacy’ and stress the importance of learning about common errors in the dialectical process before starting to build philosophical theories. It is worth quoting at some length the aim he sets for himself: I have not wanted to speak of fallacies in general, but rather of typical fallacies in philosophy. At first sight, it might appear an impossibly large task to set forth what are the sources of all dialectic illusion in philosophy. For even if there is only one truth, the possibility of error seems limitless. It is therefore highly significant for us that this is not the case. It is not the case for we can show specific types of errors of reasoning that are typically found in philosophy, i.e. that occur again and again, so that the multiple errors found in philosophy can be reduced to a definite and indeed quite small class of errors. The same sources of error, few in number, are to be found in a multitude of applications. Hence, if we investigate the sources of illusion in philosophy we arrive at a theory of dialectical error, as I would like to term this exercise—a theory that allows for an overview of possible sources of error in philosophy and that provides us with a complete systematisation of the dialectical errors that originate from these sources. (Nelson 2016: 33)

Although Nelson does not say at this point that the fallacies he is speaking of are of a linguistic nature, he also refers in the same passage to ‘the dangers that lurk in the field of dialectics’—making it clear that these are fallacies to be found in the expression of reasoning in the dialectic process, not simply in the reasoning itself, always remembering, of course, that a misunderstanding of language may cause one to reason incorrectly, not only to incorrectly express that thought. The first instances of fallacy in philosophy which he discusses are those in which the role of logic is over-estimated and consistency is taken to be the highest value of any system—Nelson points out that a good philosopher would always place truth above consistency and change his model accordingly. From here, he moves into a discussion of ‘logicism’, the Kantian synthetic a priori, and the laws of geometry, which lasts several lectures and with which we need not concern ourselves here, before finally addressing the problems of philosophical concepts and definitions. Nelson argues that, unlike geometry, philosophy should avoid what are likely to be ‘careless’ definitions because ‘before any definition is given we already have the concepts to be defined, so that an arbitrary definition always puts us into the precarious situation of confusing the concept introduced by the definition with the concept that we had before’ (2016: 123). An example is given by considering attempts at defining the concept TRUTH. However we might like to define TRUTH, in doing so we are stating that our definition is true. Do we mean that it is true according to some conventional meaning of TRUTH, in which case we are presupposing a different definition to the one we are giving, or do we mean that it is true in terms of the concept of TRUTH as it is now being defined? If it is the latter, we should

9.2 Philosophical Fallacies of Definition

139

remember that we ‘have not eliminated the concept that was previously associated with the word ‘truth’, a concept that other people still associate with the word’ (2016: 124). For Nelson this is caused by ‘the replacement of a real synthetic judgment by an apparent analytic one’; in short: ‘The fallacy consists in arbitrarily defining a concept that we are already acquainted with. When we do that, it often happens that the definition is circular in that the original concept is hidden in the definiens’ (2016: 157). The philosophical fallacy, then, is a kinsman to the persuasive definition described in the previous chapter, and a cousin too of the Naturalistic Fallacy which is discussed below. Nelson’s discussion of definitions of good, however, shows that his fallacy is not identical to Moore’s, and for the first time draws attention to the role of language in the making of it. He takes Bentham’s ([1789] 1962) definition of good as pleasure and, while he switches the example to a definition of value, shows that the word pleasure—as with others such as liking, desiring, striving for—is simply another way of saying that something is valued. The concept being defined is thus presupposed in a previous definition within the new definition. To make that plainer: what we take pleasure in is something we find to be good in some sense other than ‘good is pleasure’, so defining good as pleasure is merely to say that the new good is something which we consider to be the old good, but not some other things which we considered to be the old good, because they were not the ‘pleasure’ part of that concept. Nelson makes the observation at this point that: ‘We are in all such cases deluded by words’ (2016: 129). Nelson’s point, essentially, is that the concepts, or words, which philosophers set out to define are, in fact already understood, and cannot take as their meaning the arbitrary definition of the philosopher. He frames this as a confusion between the analytic and the synthetic, not touching on debates over the meaning of words themselves or the relation between words and the concepts they are believed to represent. It is clear, though, that the philosophers who have committed these errors, and Nelson gives many examples, have fallen into a linguistic fallacy. The Naturalistic fallacy, described by G.E Moore in his Principia Ethica, is perhaps the best known fallacy of philosophical definition. The name derives from the fact that, in the case of definitions of ‘good’, with which he is mainly concerned, a non-natural object (good) has been confused with a natural one (e.g. pleasure). However, as Moore clearly states, even without this controversial distinction, and under any different name, the fallacy remains the same: ‘It is a very simple fallacy indeed. When we say that an orange is yellow, we do not think our statement binds us to hold that “orange” means nothing else than “yellow,” or that nothing can be yellow but an orange’ (Moore 1903, ch.1, §12). Notwithstanding the fact that oranges are clearly orange and rarely yellow, the point of the fallacy Moore is drawing attention to is very simple: when we predicate a complement of a particular subject, we do not state that they are identical things. We cannot reverse the order as we can with a sign of equivalence: a = b means b = a, but a is b, does not imply that b is a, because although the word ‘is’ can be used to mean identity, it generally isn’t. This is pointed out by Wittgenstein in the Tractatus: ‘the word ‘is’ figures as the copula, as a

140

9 Linguistic Fallacies in Philosophy

sign for identity, and as an expression for existence […] In this way the most fundamental confusions are easily produced (the whole of philosophy is full of them)’ (3.323–3.234; 1974: 15–16). Both the difference with Nelson’s fallacy and the similarity should be readily apparent. In both cases the philosopher comes to grief by trying to define one existing word as really meaning something else. The actual error made in so-doing is not the same though: in Nelson’s case the error is to presuppose the prior understanding of the term in the new definition, whereas for Moore it is actually reducible to a confusion over the two uses of the verb ‘to be’; whether one thing is the same as another thing, or whether the second is merely a characteristic of the first. It is also true that in both cases the way the fallacy is introduced hides its purely linguistic nature. Nelson sees the errors he describes as a result of a failure to recognise synthetic a priori judgements and Moore suggests that his come from not recognising that ‘good’ is a simple which has no further definable constituents. Debate over their fallacies may be easily side-tracked into a debate over these ideas, but the essence of the error for both is a linguistic one. One is a form of persuasive definition and the other, at root, the result of ambiguity. A further, if apparently mundane, danger is highlighted by Arne Naess, who notes that ‘it is a common malpractice to introduce formally a prescriptive definition and then quickly forget to follow it or fail to follow it consistently’ (1966: 46). This can be particularly problematic when one wants to discuss the works of others who have defined and used the word differently: it may soon become hard for the reader to tell to whose definition a certain item answers, and he might begin to wonder if the author is in a similar strait. The scope for equivocations brought about this way, both by defining authors and those who would interpret them, is wide. An important general point we can take from this about the nature of language is that no word can ever be defined by one other word unless the other word has exactly the same meaning as the first. When such rare cases can be found, they are mere tautologies and do not tell us anything new, unless someone did not know one of the words previously. If I know the two words, say good and pleasure, or truth and beauty, or God and love, then I know that the one is not identical with the other. No knowledge can ever be created in this way. It is to misunderstanding the nature of meaning to think that it could be otherwise. Some of the examples cited by Nelson are such obvious cases of circular definition that one wonders how intelligent people ever came to write them down, and still more that others saw fit to give credence to their views; and, as the Naturalistic fallacy is presented by Moore, it is hard to imagine any serious thinker actually committing such an error. So, why have so many thinkers tried to define the good and other concepts which are perfectly well used in everyday discourse, as though there were some great mystery about their meaning? One explanation is that there is a kind of upward pressure from the common man who expects the intellectuals of his tribe to answer certain vital questions which he asks himself, but cannot even see what path he might take towards a solution. The common man, I suggest, wants to know what is true and what is good: the philosopher cannot answer these questions, so instead turns his attention to explaining what is truth and what is goodness. The

9.2 Philosophical Fallacies of Definition

141

common man, however, understands these terms perfectly well. He wants to know which statements, which claims, which doctrines, are true and which are falsehoods; if he didn’t already understand what truth and falsehood are, he wouldn’t be asking the question. Thus, any definition of truth is always delivered to an audience which already knows what it means to be true and can then consider at their leisure whether the new definition is true or not, based on their own knowledge. Any definition of good will always be judged as to whether that which meets its requirements is, actually, good or not, understood as fitting the original meaning of the word. One obvious objection to the ‘good is pleasure’ theory is that some acts of evil may bring pleasure to the perpetrators—and acts of evil are not good. Many philosophers, it seems, have not realised that even the simplest of their countrymen speak and understand the language of that country. They do not need to have words defined for them.

9.3 The Fetish of Language Another possible cause of this confusion among the cleverest is the almost occult power which language holds over those who most think it their servant. To be an intellectual is to be a master of language, but it often means to place the greatest faith in the ability of language to both express all that needs to be said, and to reflect the world just as it really is. This leads to the tendency to think that if a word exists, then something must be its referent, yet it is not for nothing that Naess reminds us ‘that the existence of some concept term in no ways guarantees that something falls under the concept’ (1966: 67). There is a whiff of Anselm’s Ontological argument about this way of thinking: the simple way to refute that argument is to deny that one can conceive of a perfect being, whether existing or not existing. In the same vein, I may reject any definition of ‘the good’ on the basis that, although I understand how the phrase is commonly used, I cannot conceive of anything that it could have as a referent. It is thoughts such as these which make me wary of distinctions between words and concepts. I can understand ‘God’ as a word, with a use in a community of speakers, but I cannot conceive of a separate but matching concept, only more words. This is also a de facto rejection of Platonism: it isn’t a question of whether a ‘form’ of the concept of ‘chair’ exists in reality or only in one’s mind, it is clear to me that no archetypal chair, nor any idea of one, can exist anywhere. There is no concept of ‘chair’ which is any different from the word ‘chair’, a word I use and understand without the slightest reference in my mind to any gold standard. The belief that when we use words we are referring to some standard of reference is simply false. Another way of looking at examples of the Naturalistic fallacy is to see them as akin to the fallacy of ‘paronymous words’ described in the previous chapter. In this case there is an apparent misunderstanding of the difference in meaning between members of a related noun/adjective pair. When I see that pleasure (noun) is good (adjective), I have no right to move to either pleasant (adjective) is good (adjective), nor to pleasure (noun) is (the) good (noun), and certainly not to good (noun) is

142

9 Linguistic Fallacies in Philosophy

pleasure (noun), or good (noun) is pleasant (adjective). Each statement is different: all may be true, but they do not follow from one another. Indeed, one may wonder if the very existence of nouns such as ‘good’ is not a result of purely grammatical processes. Adjectives can be turned into nouns, and so we take a common adjective of approval ‘good’ and create a noun ‘(the) good’ to mean, what exactly? It is a fascinating, and in some ways terrifying, idea that a large part of our vocabulary has slipped into use in order to satisfy the demands of grammatical categories, rather than to describe or explain phenomena in the experience of our lives. There is an echo of the concerns discussed in Chapter 1 here, that language may be shaping thought, even as thought was first creating language. This is the essence of Berkeley’s complaint about all ‘abstracts’. The fact that an adjective can be used successfully to identify actions, events and objects in a certain way does not mean that there is an equally intelligible overriding abstract concept in which they all somehow partake. That is the simple error which runs through much of philosophy from Plato onwards. Berkeley’s point was mentioned briefly above, but I shall revisit it in more detail here. In the Introduction to his Principles of Human Knowledge, he observes that the notion of ‘abstract general ideas’ is the most influential false principle afflicting the minds of ‘speculative men’, and lays much of the blame at the feet of the Schoolmen. He then comes, in section 18, ‘to consider the source of this prevailing notion, and that seems to me to be language’. This is proven by the fact that it is admitted by ‘the ablest patrons of abstract ideas, who acknowledge that they are made in order to naming; from which it is a clear consequence, that if there had been no such thing as speech or universal signs, there never had been any thought of abstraction’ (Berkeley 1988: 47). This statement makes a direct link back to the first part of this volume and considerations of the role of language in shaping our thinking. It seems that because we can give names, we do so, and then because there is a name, we assume that it is the name of something: ‘Whereas, in truth, there is no such thing as one precise and definite signification annexed to any general name’. This is a direct reaction to a passage in Locke’s Essay Concerning Human Understanding, the target for much of the work. Locke makes it clear in Bk III, ch.vi, that abstract ideas are the creation of men through their giving of names to collections of ideas, and that the giving of a new name creates a new abstract: ‘Nature makes many particular things, which do agree one with another, in many sensible qualities […] but ‘tis not this real Essence that distinguishes them into Species, ‘tis Men, who […] range them into sorts, in order to their naming, for the convenience of comprehensive signs’ (Locke 1975: 462). He uses the example of the division of timepieces into watches and clocks noting that ‘the two are but one species, to those who have but one name for them’ (1975: 463). For Berkeley, there is an inconsistency here in that Locke has much to say on abstract general ideas and the process of abstraction, even though he acknowledges that such ideas come about purely as a result of the desire to name things. Berkeley cites his rival’s description of the abstract of a triangle: ‘it must be neither Oblique, nor Rectangle, neither Equilateral, Equicrural, nor Scalenon; but all and none of these at once’ (Locke 1975: 596) and suggests that he cannot form an idea of such a triangle, and he is right, no-one could. It seems unlikely that Locke would not have agreed that such a triangle could not be imagined, but he might have pointed out that the abstract

9.3 The Fetish of Language

143

wasn’t a triangle; it was a collection of properties defining triangularity. That passage from Locke, however, does suggest that he has an impossible triangle in mind, and this is an example of the danger such notions introduce. The triangle example is a good one because triangles are real things which we can see and identify with ease, yet any archetypal triangle would have to have an impossible range of possibilities. We can see then that the abstract of the triangle does not exist as a thing which can be imagined as really existing: the word means only a three-sided shape, in the same way that a bachelor means an unmarried man, and we can no more imagine a triangle with only that quality than we can imagine a man with only the absence of a wife. With other words this is less obvious. As we have seen, when the concept of ‘the good’ is abstracted from individual instances of ‘things which are called good’ what we have is a rough grouping of qualities, some of them contradicting each other, not a unified comprehensible concept. When a philosopher comes along and tries to define the good as being something consistent, something represented in its entirety by something really existing, he is, in effect, drawing a right-angled isosceles triangle.

9.4 A Critique of Language There can be little doubt that the thinker whose ideas on language, and its unfortunate effects on the minds of philosophers, have most permeated the wider cultural consciousness is Ludwig Wittgenstein; and while scholars make much of the change in his views between the writing of the Tractatus and of the Philosophical Investigations, what is constant is the centrality of language to his understanding of both the methods and the purpose of philosophy. There is no space here, of course, to discuss the Ideal language of Logical Atomism, the Verifiable Language of Logical Positivism, or even the Ordinary Language, of the later Wittgenstein and the Oxford school, nor the theories of truth which led to them, save to say that all do see philosophical truth and the search for it as deeply entwined with language. Indeed, this is true to such an extent that it becomes difficult to separate philosophical truth from the language in which it is expressed. It might be an extreme view to suggest that all philosophical truths, assuming such things exist, simply are truths about language; and that, therefore, discussion of them is certain to be fraught with errors brought on by misunderstanding and unfortunate expression. It is a view, however, with which I have a great deal of sympathy, and one which I believe is supported by many of the points made about language by Wittgenstein at all stages of his career. This strand of thought must, sadly, remain no more than a faint undercurrent in the present study. The following paragraphs will consider Wittgenstein’s comments on language only as they are directly concerned with the process of philosophical debate. In the Tractatus, Wittgenstein famously writes: ‘All philosophy is a “critique of language”’ before adding in parenthesis, ‘though not in Mauthner’s sense’ (4.0031, 1974: 19). In order to understand Wittgenstein’s sense, then, we need to quickly deal with Mauthner’s. In a three volume work (1901–1902), Mauthner puts forward a

144

9 Linguistic Fallacies in Philosophy

view of language which suggests that it has little relation to reality. According to Gershon Weiler, Mauthner’s ‘main and often repeated thesis is that “Knowledge of the world through language is impossible, that there is no science of the world and that language is an insufficient instrument for knowledge” (B. 2nd edn. i, p. xl)’ (Weiler 1958: 81). Language is a reflection only of our own sense experience of the world: senses which are accidental in their evolution. This is obviously in opposition to the view found in the Tractatus, where ‘A proposition is a picture of reality’ (4.01). For Mauthner: ‘As it is futile to try to construct a perfectly logical language, so it is futile to ask for the logical grounds of our ordinary usage of language. […] If the word has no meaning in isolation, then, of course, the question ‘What does X (standing for a word) mean?’ becomes unjustified. M[authner] regards it “as a piece of superstition and mental weakness [...] that because there is a word, it must be a word for something, that because a word exists there must exist something real that corresponds to that word” (B. 2nd edn. i. 159)’ (Weiler 1958: 84). Mauthner’s critique, then, when applied to philosophical practice, is similar to that of Berkeley. The existence of words does not imply the existence of something in the real world to which they refer. His view of the role of philosophy is that ‘the most important business of philosophy is the critique of language. No linguistic expression can be taken as representative of reality, so the philosopher’s task is to drive home to us this idea and to free us from the spell of language. This will be the self-criticism of philosophy through the criticism of language’ (Weiler 1958: 85). Any philosopher who attempts to express truths in language is guilty of a misunderstanding of what language is. Mauthner recommends that men return to silence; yet, like Lao Tzu before him, he has only language with which to make these observations. Wittgenstein’s critique of language is different because he believes, at this stage at least, that a properly constructed, logical language can describe the world in its entirety. This language would be free of ambiguity and other confusions of natural language (see 3.322–3.325). The stages of his argument can be stated easily: 1.2 The world divides into facts. 2.141 A picture is a fact. 2.19 Logical pictures can depict the world. 3 A logical picture of facts is a thought. 3.1 In a proposition a thought finds an expression that can be perceived by the senses. So propositions, in an ideal language, are expressions of thoughts, which are themselves logical pictures of the facts which make up the world. The critique of language, then, is that so much of it is not logical, and, therefore, since ‘What a picture represents is its sense’ (2.221), it is nonsense. Wittgenstein stresses in his preface that he’s not much interested in whether other philosophers have thought similarly about language and he gives no specific examples of philosophical nonsense of this sort. He does, however, make his opinion on large swaths of philosophy clear. In point 6.41 he notes that no value exists in this world of facts which can be represented by propositions, so that if there is value ‘it must lie outside the whole sphere of what happens and is the case.’ The corollary of this

9.4 A Critique of Language

145

(6.42) is: ‘So too it is impossible for there to be propositions of ethics. Propositions can express nothing that is higher’ and (6.421) ‘It is clear that ethics cannot be put into words’: as Lao Tzu has said. It seems, then, that every philosopher who has tried to discuss ethics or any topics concerning moral values has been guilty of talking nonsense. Indeed, according to Wittgenstein, the proper role of philosophy is only to state propositions of natural science and ‘whenever someone else wanted to say something metaphysical, to demonstrate to him that he had failed to give a meaning to certain signs in his propositions’ (6.53). Naturally, it is not easy to see how this conclusion could be employed in our ultimate goal of the linguistic analysis and assessment of argumentation. We can hardly reject any argument that contains propositions not concerned with scientific fact. We can, however, take the more general point that some sentences may attempt to express in language that which cannot be meaningfully said, or at least, that which cannot sensibly be argued about: I cannot argue with a man about his own mystical experiences, and I may consider as fallacious any argument in which he attempts to use them as evidence on the grounds that his words have no real meaning. As Wittgenstein famously concludes: ‘What we cannot speak about we must pass over in silence’ (7). Mauthner and Wittgenstein, then, see language differently: for one it can only describe the personal, and the true state of affairs can be understood only mystically, hence silence is advisable; for the other, only propositions describing facts are meaningful, and the personal experience of existence can only be expressed mystically, hence silence is advisable. For both, the necessity to set forth their ideas in language creates a difficulty: Mauthner doesn’t see language as a capable of being a conveyor of truth, and Wittgenstein is forced to admit that anyone who has understood his work knows that what he has said is itself metaphysical and thus nonsensical. Whether by coincidence or not, both men use a metaphor of language as a ladder which one may climb, but the rungs of which should be destroyed once they have been used (Nájera 2007). This remarkable reflection of the other’s views is fascinating, and becomes more so when we look at how Wittgenstein’s thought later developed: for Mauthner, meaning was use, not reference to something existing, an idea which Wittgenstein would make famous, and for which he became widely known. The differences between early and late Wittgenstein can be over-played. Although they clearly exist, the centrality of the role of language in philosophical endeavour is a constant and is seen throughout the notebooks he kept while preparing for the Philosophical Investigations. There is, through the whole course of his work, a feeling that proper language use is the essence of good philosophy, and while his views on how language should properly be employed may have altered, this makes him a constant friend of our project. There are two particular strands of worry over language use which are developed by Wittgenstein in the later period. Neither of them is especially original and they provide a rounded summary of the problems philosophers have had with other philosophers over their use of language. The first is a return to the problem of the existence of a word being taken to imply the existence of its referent. This is the subject of the opening section of the Blue Book. He writes: ‘The questions “What

146

9 Linguistic Fallacies in Philosophy

is length?”, “What is meaning?”, “What is the number one?” etc. produce in us a mental cramp. We feel that we can’t point to anything in reply to them and yet ought to point to something’ (1958: 1). That he considers this an important problem in the history of the subject is confirmed in the following sentence in parenthesis: ‘We are up against one of the great sources of philosophical bewilderment’. An example of this is found when we consider ascriptions of mental states and the space within which they may be said to exist: a space generally called ‘the mind’. Wittgenstein writes: ‘when we are worried about the nature of thinking, the puzzlement which we strongly interpret to be one of the nature of a medium, is a puzzlement caused by the mystifying use of our language. This kind of mistake recurs again and again in philosophy’ (1958: 6). This is nicely summarised by Morris Engel: ‘the existence of the word ‘mind’ leads us to look for a thing corresponding to it, and then to assign to it […] all sorts of occult powers’ (Engel 1971: 17). These ‘occult powers’ comfortably explain the way in which the word is used, but are themselves beyond all explanation. The second way in which misconceptions of language may lead to bad philosophy is hinted at through the warnings Wittgenstein offers to the philosopher against the temptation to take language and, as Engel puts it, ‘revise it in ways more congenial to him’ (Engel 1971: 42). In the Philosophical Investigations, Wittgenstein writes: ‘Philosophy can in no way interfere with the actual use of language’ (Wittgenstein 1953: §124, 49e) and then: ‘It is not our aim to refute or complete the system of rules for the use of words in unheard-of ways’ (§133, 51e), which, despite the lack of any examples, does seem to imply that that is what philosophers had thitherto been attempting to do. A special form of this type of behaviour is suggested in Toulmin’s argument for making distinctions between different types of premises. In a section dealing with ‘Universal Premisses’, he considers the popular form of statement ‘All A’s are B’s’ and the debate over whether making it commits one to accepting the existence of some A’s. There is a sort of ambiguity in such a premise since it is unclear whether it is a factual record of what has been observed of all A’s or whether it relates directly to the nature of being A, that it inherently means being B too. In the latter case, it matters not if nothing is, nor has ever been, A. Since a distinction between types of premise is not usually found, we cannot be sure whether an existential commitment is being made or not when all A’s are stated to be B: ‘Some of the statements which logicians represent in this rather crude form do have such implications; others do not’ (Toulmin 2003: 107). This point can be expanded to a more general one about the way arguments are written down in order to receive logical analysis. While tidying up and formalising statements might appear to make those arguments more logically expressed, and certainly makes their analysis easier, Toulmin’s example shows that simplified forms may introduce ambiguity by discarding more nuanced elements of the original vernacular. For the form ‘All A’s are B’s’ occurs in practical argument much less than one would suppose from logic text-books: indeed, a great deal of effort has to be expended in order to train students in ways of rephrasing in this special form the idiomatic statements to which they

9.4 A Critique of Language

147

are already accustomed, thereby making these idiomatic utterances apparently amenable to traditional syllogistic analysis. There is no need, in complaining of this, to argue that idiom is sacrosanct, or provides by itself understanding of a kind we could not have had before. Nevertheless, in our normal ways of expressing ourselves, one will find many points of idiom which can serve as very definite clues, and are capable in this case of leading us in the right direction. (Toulmin 2003: 108)

It is hard not to think that the writers of logic textbooks believe their versions, cleansed as they are of the messy vagaries of everyday idiom, to be superior in representing the ‘real’ claims of the argument. Formalisation, however, has to mean simplification, and there is danger inherent in that. Let us remember too, that that danger is also present for the informal logician who is required to rephrase a natural language argument in order to assess it. To conclude, Wittgenstein can be said, in his typically plain yet inexplicit style, to have raised all three of the great philosophical fallacies of language: the language fetishism of Berkeley and Mauthner, which would have it that language so well reflects the world that that which it names must be found therein; the language revisionism of Nelson and Moore, by which philosophers seek to better explain the world by redefining the words in which it is described to suit their own explanations; and, finally, the misunderstanding of the very nature of language, what it is capable of expressing and on what we ought to preserve a careful silence. The paradox at the heart of philosophical discourse then is this: if we attempt to describe the truth about the world using words we know in grammatical forms we recognise which allow of public argument and discussion, we are limited by the conventional meanings of those words, and any attempt at redefinition is likely to be fallacious, because it is defining something different from the meaning which the word originally had. However, if, like the Taoist, we try to describe a new concept, a word which is not commonly understood, then we run up against the inadequacy of the existing vocabulary to describe it. A mere compound of two other words is not a problem, so long as those words do not have entirely contradictory semantics, but something which goes beyond the experiences of men which have already been manifested in language, that, by definition, has no language to describe it, must forever remain opaque and mystical. It is hard enough for two speakers of a language to be sure that the word they use is understood by others in the same way that it was meant when that word is common and born of direct experience; for two mystics discussing that which cannot be discussed, it is a literal impossibility. The temptation then is to take the vocabulary of the mystic and claim that it refers to the world in the way that names for objects do. The process by which meaning and reference take place may be controversial for philosophers of language, but nobody doubts that some words refer to things which actually are in our world and others refer to things which are not. Still others are names which are useful to us in our discourse, but are not linked to anything which might be considered real and can be described in all its qualities. These words are often taken to mean something real, because they are useful and can be understood, but a word’s being understood does not mean that its referent can be.

148

9 Linguistic Fallacies in Philosophy

9.5 Vagueness and Incompatibility The final section of this chapter will contain a brief word on two types of linguistic fallacy found in philosophy, but not related to this paradox of language just described. The first, and most easily dealt with, is the general charge of vagueness. That much philosophy has been written in prose so dense as to be impenetrable is, I think, a fact that few would dispute; that that very difficulty in being understood may be an advantage for the philosopher in that his works may be picked over for generations to come, and subjected to as many variations of interpretation as the creativity of the human imagination can invent is a sad truth, telling us much about professional philosophers and their indulgence of obscurity of expression; and, that much of the field known as Continental philosophy delights in, celebrates and propagates the use of words without sense in a pseudo-intellectual mystical ceremony of mutual affirmation should surprise no-one with a reasonable acquaintance of human nature. Given these facts, further examination of the rampant obscurantism in philosophical writing is unnecessary here. Let it be noted simply that all the sins of vagueness are to be found amongst the greatest philosophers of old and the trendiest of the new. In fairness, this can partly be blamed on the territory: when one is attempting to explain the inexplicable, clarity is bound to be elusive; when one is recreating language, misunderstanding is to be expected; and when one is surreptitiously shifting from ‘is’ to ‘ought’, arguing in a circle, or hiding one’s own weak understanding, imperspicuity is perhaps advisable. Finally, I offer an example of the type of fallacy which I identified in the previous chapter under the name of Semantic Incompatibility. It is, perhaps, surprising, given its prominent place in the history of the study of Ethics, that Utilitarianism proves to be such a fecund source of linguistic fallacies; every aspect of the theory, it seems, relies on one or other of the typical misunderstandings and abuses of language, any one of which ought to have been enough to see it forgotten about very soon after it was broadcast. We have already seen how the move to define good as pleasure falls foul of both the Philosophical fallacy of Nelson and the Naturalistic fallacy of Moore. Yet even if we were to ignore this weakness, and even if we were to imagine for a moment a world in which the word ‘Utility’ had a sensible meaning other than the mundane ‘usefulness’ which most of us would ascribe to it, the arguments offered in its favour employ words in ways which make them simply incompatible with each other. The suggestion that one should act to increase pleasure in society is questionable, but not itself fallacious. Jeremy Bentham, however, offers a description of the process by which one should choose how to act, urging that we: ‘Sum up all the values of all the pleasures on the one side, and those of all the pains on the other’ for the individual and then again sum those numbers for all individuals and ‘take the balance’ (Bentham 1962: 66). One possible criticism of this is to state the practical difficulties of making such measurements, but this is unnecessary: the phrase ‘sum the values of the pleasures’ contains a clash of semantics and is essentially

9.5 Vagueness and Incompatibility

149

meaningless. Only numbers can be summed, and pain and pleasure are not numerical. The odd collocation of words only gets worse when we move to considerations of average utility. John Rawls explains that on this view ‘institutions are set up so as to maximize the percentage weighted sum of the expectations of representative individuals. To compute this sum we multiply expectations by the fraction of society at the corresponding position’ (1972: 162). Expectations are defined by Rawls as indicators of ‘life prospects’ (1972: 64). The language of mathematics is mixed with the language of value to produce a seductive, but illusory, simplicity. Argumentation theorists are still seeking ways to ‘sum’ arguments in order to determine their combined strength; in formal argumentation, arguments are, rather arbitrarily, given numerical values so that they can be added together. As unsatisfactory as that procedure may be, there is really no other way. If reasons cannot be added together, what hope for the summing of pleasures and the multiplication of expectations? It is easy to understand the temptation to talk in this way. Mathematics is a very useful tool providing very clear answers. That moral philosophers who struggle with conflicting rights and incompatible duties would appreciate the clarity of a spreadsheet is unsurprising, though it seems to miss the very essence of what ethical thought is. So too that formal theorists of informal argument are eager to feed numbers into their Dung frameworks or Bayesian theorem and so reduce human reasoning to the simplicity of machine computation, and yet the provenance of these values remains opaque. In conclusion, I believe the concern shown by philosophers over the misuse of language in their field and the examples cited of such errors and confusions support my suggestion that, at this level of debate, falling into linguistic fallacy is both likely and calamitous. Philosophers are duty bound to attempt explanations of that which language is not suited to explain. They are forced then to twist language and redefine its shape and meaning, or to accept it as it is and wallow in the illogical and confused mire it depicts. I return then to the question raised at the beginning of this chapter: is all philosophy a mere squabble over words with no hope of shedding real light on the world or the state of man? That would be a drastic conclusion to reach. Certainly, philosophers are in a difficult position whenever they try to build any kind of a system, as at every step they are fighting against the language whilst simultaneously working with it. However, the practice of philosophy as a system of analysis, taking language on its own terms, is secure. Argumentation is on its surest ground when it is assessing arguments and seeking errors within them. The final part of this work will be dedicated to a description of how that analysis, in particular analysis of the language employed in argument, can best be carried out.

Part IV

Analysis

He draweth out the thread of his verbosity finer than the staple of his argument William Shakespeare, Love’s Labour’s Lost, V.1.

Chapter 10

Analysis of Arguments

The preceding chapters have covered a great deal of ground in their attempts to explore how language and argument are related. It was inevitable that in order to properly understand that relation, arguments themselves and the process, or processes, of argumentation in which they are employed would also need to be dissected and described. Interesting as that may have been in and of itself, the core purpose of this study was to develop a better understanding of how the linguistic element of arguments can be analysed and assessed, and to use that improved understanding in the development of tools with that function: in particular an informal semantics of argumentation. What remains, then, of this volume is dedicated to the task of assessing arguments. This will involve setting out a comprehensive procedure for argument analysis and assessment, in accordance with the definition of argumentation developed in earlier chapters, and a more detailed framework for the analysis of the language of argumentation. These procedures will then be shown in action, put to work on a number of texts drawn from real-world sources.

10.1 Three Categories of Assessment One of the great advantages of this type of systematic assessment is that if an argument is found to go wrong, it must go wrong at a particular stage of the analysis. This means that failure at each stage may be labelled as a particular kind of fallacy: rejection during the linguistic analysis means that there is a linguistic fallacy, and if that failure occurs during the clarity analysis, then a fallacy of vagueness has been committed. In this way the fallacies are ordered systematically. No longer a rag-tag collection of errors loosely placed in different groups at the fancy of each author commenting upon them, ripe to be disputed; but occurring in defined ways at a defined point in the process of analysis. This is a serendipitous by-product of the attempt at schematising

© Springer Nature Switzerland AG 2021 M. Hinton, Evaluating the Language of Argument, Argumentation Library 37, https://doi.org/10.1007/978-3-030-61694-6_10

153

154

10 Analysis of Arguments

the assessment process. Viewed from this perspective, it is rather surprising that such approaches have not been widely employed before. All of that, obviously, presupposes that the term fallacy can be applied to any errors detected in the evaluation procedure, and only to such errors. This is a supposition I intend to make, and it has several corollaries. Firstly, since the full procedure is only applied to what has already been identified as an argument, fallacies are considered as things which can go wrong with arguments, but not as arguments in themselves. Earlier definitions have perhaps equivocated on the use of the word to refer to types and instances: here, a fallacy is a type of error; an instance of one is a fallacious argument. Also, they are things which can go wrong with arguments, not with argumentative discourse in general: where one refuses to offer any argument, one cannot commit a fallacy. This is a step closer to the ‘traditional’ conception and away from the pragma-dialectical one. That step, however, is only a very short one. Only the E for erroneous is required from Woods’s EAUI list: the idea that a fallacy must be common and must be likely to deceive is dropped. This move circumvents many of the problems with the traditional theory, while maintaining the notion that fallaciousness is a property of arguments rather than of discourse. A further consequence is that the traditional naming of fallacies becomes somewhat irrelevant. When assessing a piece of argumentation, the analyst no longer looks for structures analogous to those in certain named and established patterns of error, but simply applies the evaluation procedure until an error is uncovered. This error may equate to an instance of a named fallacy and be recognised as such, or it may not. Either way, the reason for rejecting an argument is its failure at one level of evaluation, not its similarity to a textbook example of a ‘logical fallacy’. That is not to say that the traditional fallacy names have no place in the everyday description of arguments or in the teaching of critical thinking, but it is to deny them a place in a systematic account of argument error. Table 10.1 gives a general overview of how such a systematic typology would look on the basis of the assessment procedures introduced in the remaining chapters. Each of the three sections is explained in greater detail in Chapter 11, but since the focus of this work is on language and argument, the Process and Reasoning categories are not fully expanded, and contain only a few illustrative examples in the later columns. Some use is made of traditional fallacy names to assist the reader. The categorisation in the table is taken as far as third order sub-types, but could certainly be taken deeper if finer divisions were considered necessary. Before beginning the process of showing how arguments can be analysed according to the conceptions of this work, however, there are three topics which need to be discussed in relation to that analysis: meta-argument, reconstruction, and normativity. Meta-argument is an awkward concept since it is inherent in the activity of arguing that many arguments are arguments about other arguments, meaning that argument analysis cannot stand as a separate procedure, independent from argument formation. Some discussion of this point will be found in the section below. The section after that will tackle the difficult area of argument reconstruction. The fact is that very few actual arguments contain a full logical structure which can

10.1 Three Categories of Assessment

155

Table 10.1 A partial typology of fallacies Type of fallacy

Sub-type first order

Sub-type second order

Sub-type third order

Process

Relevance

Straw man

Invented position Misrepresented position

Begging the question Appropriateness

Burden of proof

Premise

False data

Inadmissible evidence Reasoning

Unknowable data Lever

Logical error

Denying the antecedent

Weak inference

Anecdotal evidence

Affirming the consequent Language

Vagueness

Precision

Ambiguity

Equivocation

Obfuscation Ambiguous claim Amphiboly Compatibility Philosophical

Definitional

Nelson’s fallacy Persuasive definition Naturalistic fallacy

Concept

Abstract concept Cognate Parts of speech Mysticism

be examined, and accepted or rejected without any expansion: one premise is often missing, sometimes many premises are missing, sometimes only the premises are given and we are left to infer the conclusion ourselves. This is a particular problem for any linguistic assessment, since the analyst is left with no option but to put words into the arguer’s mouth, words for which that person may deny responsibility. Thirdly, the issue of normativity cannot be avoided any longer. So far, a certain attitude to the rights and wrongs of argumentation has been assumed without question, but that attitude must now be made plain, and justified. An argument can be analysed from a non-normative perspective, but the fundamental question for most scholars is not what is an argument so much as what is a good argument. The short answer to that question will ‘be anything that is an argument and is not bad’. The process of analysis will seek reasons for the rejection of arguments and the title of good argument will be awarded by default on the principle of ‘last man standing’ to whatever is not rejected.

156

10 Analysis of Arguments

Finally, as a preliminary to the general assessment scheme, I briefly discuss how the structure and methodology of this form of evaluation relates to other systems for judging and describing argumentative discourse: in particular, the rule-based pragma-dialectical approach.

10.2 Meta-Argument If meta-argument is understood as argument about argument, then such a good deal of argument is meta-argument that it seems unnecessary, and perhaps impossible, to separate the two. Certainly, using the term meta-argument to refer to the normal to and fro of defending, rejecting, clarifying and commenting upon before ultimately accepting or rejecting arguments offered in everyday discourse would leave the term argument with very little territory for itself. For this reason, responses to argument which are part of the process of argumentation should not be regarded as metaargument, just more argument. However, that leaves a lot which can be considered meta-argument, and I shall suggest that the assessment schemes outlined below belong within that classification. For Daniel Cohen (2001), meta-arguments are those which we can employ to justify accepting or rejecting an argument. These meta-arguments are not normally articulated: ‘It is quite within proper argumentative practice that we do not routinely use the meta-argument for accepting good arguments’ (2001: 80) since they might lead to an infinite regress, yet they lie behind our reasoning. There is something unsatisfactory about this account; the possible objections Cohen raises never seem to be properly quashed, but there are two points of interest. Firstly, meta-arguments are principles of acceptance or rejection, not comments on individual arguments with which we are faced, and secondly, ‘these reinforcing arguments are at home in self-reflective deliberation […] the context in which the distractions created by the competitive and social aspects of argumentation are largely absent’ (2001: 81). Both of these characteristics are reflected in the assessment schemes below: they set out general principles for rejecting arguments and they are not suited to the actual hurly-burly of real-time debate, since they emphasise close and careful textual analysis.1 The best-known writer on meta-argumentation is Maurice Finocchiaro. He has written a number of works on the subject, notably his (2007) paper and a book (2013). In his simplest definition he describes ‘a meta-argument as an argument about one or more arguments’ (2007: 253). In the longer work he gives three types of metaargument: argument analysis, self-reflective argument, and argumentation theory. None of these quite matches Cohen’s arguments for accepting arguments, but they are in agreement on the important role of meta-argument in considering one’s own 1I

would suggest, however, that thorough acquaintance with the procedures would be of great use in enabling the fast recognition and exposure of error in arguments offered in a social discourse environment.

10.2 Meta-Argument

157

arguments. This is not something which has been discussed in this volume so far: although solo argumentation was one of the modes considered in Chapter 5, and it is an interesting question to what extent examining one’s own arguments resembles responding to those put forward by others. There doesn’t seem to be any reason why the procedures below could not be applied to what one has said or thought oneself, in a spirit of self-examination. It is, I think, safe to say that, at least, some arguments about argumentation theory can be called meta-arguments, although they remain arguments nonetheless, and so the contents of this book as well as others in the field, deserve the label meta-argument. The most difficult issue with the term comes from the first of Finocchiaro’s categories, that of argument analysis, as it can be problematic to see the line between analysis and counterargument. Anthony Blair worries that on such a definition ‘defending a claim against argued-for objections is meta-argument’ (2014: 237): though one is merely supporting one’s original position, one does so by commenting on the argument against it, that is, offering an argument about an argument. This is why the distinction I made at the beginning of this section is important: argumentative discourse which is aimed at analysing, interpreting and assessing arguments outside the process of argumentation can be considered meta-argument, but responses to arguments, even when based on meta-argumentational principles, are simply counterarguments and part of argumentation, not meta-argumentation. Thus, two lawyers discussing the admissibility of a piece of evidence are arguing, two legal theorists discussing the rules of admissibility of evidence in legal argument are engaged in meta-argument. The assessment procedures, then, can be said to contain a representation of a number of meta-arguments which support the rejection of an argument at any particular point in the process of analysis, while the earlier, theoretical chapters of this work stand as the backing for those meta-arguments. The prefix ‘meta’ however is always relative and a critic reviewing the arguments I have put forth, and the construction of the assessment schemes, may raise meta-arguments against them, not by simply showing them to be wrong, but by questioning the principles of reasoning and the understanding of argumentation upon which they have been made. One man’s meta-argument is another man’s argument to be ‘meta’ about.

10.3 Reconstruction Any attempt at the analysis of argumentative natural language discourse will inevitably require that a degree of reconstruction take place. It is a simple fact of human language use that points are not made with clearly delineated premises leading to an explicitly stated claim. As speakers, we generally follow the economy principle (Grice 1975) and don’t say aloud that which is obvious to everybody or which we believe is already accepted by the present company. This often means that what Toulmin calls warrant premises are left implicit,2 as described in Chapter 4, 2 For

an interesting recent approach to identifying implicit warrants, see Habernal et al. (2018).

158

10 Analysis of Arguments

and conclusions too may not need to be stated. It is not then a fault of an argument, at least in everyday conversation, that it is somehow incomplete in its expression; but it is an inconvenience to anyone who wishes to analyse that argument carefully. In an early paper, van Eemeren (1986) discusses the differences between a ‘normative reconstruction’ and a simple description of a piece of discourse, identifying three procedures which are necessary: Selection, which also means deletion; Completion, which involves addition; and Arrangement, or permutation.3 These seem to catch the essence of the task: one must take the original text and select which parts of it are essential to the argument, and which ‘data are left aside as immaterial’ (van Eemeren 1986: 10) and are thus deleted. This may be obvious, in the case of repetitions, for instance, or require a considered judgement on the relevance of some items. Since what is expressed is unlikely to form a full argument, one must then complete it. This involves two separate processes: the making explicit of that which was left implicit, and ‘adding elements whose presence has to be assumed in the case of a fully-fledged dispute’ (1986: 10). This second is vital and should not be forgotten: one cannot judge the relevance of an argument if one does not know the argumentative context in which it is being made. Thirdly, this all now needs to be properly arranged: ‘normative reconstruction of a dispute need not directly reflect the linear course of events’ (1986: 11), and indeed, given the way natural language users tend to place their premises both before and after their conclusions, it may be a rare occasion when it does. To this list is later added Substitution, which ‘amounts to reformulating in the analysis in unequivocal and clear paraphrases those standpoints and other crucial elements in the discourse whose function in the resolution process would otherwise be unnecessarily opaque due to ambiguous or vague formulations’ (van Eemeren et al. 2014: 536). This activity is subtly different from the others in that the first three are necessary to establish that there is an argument contained within the text and to see how it is structured (this is the reconstruction which takes place after initial analysis in the scheme below), while the fourth is a process of improving or making clearer the argument (which is labelled ‘reform’ in the diagram below, and occurs after the linguistic stage of analysis). These four processes, however essential, are all obviously open to interpretation. Mark Aakhus points out that while reconstruction is a normal part of argument—‘When one person holds another person accountable for saying, intending, or meaning such and such, what was said has been reconstructed as having argumentative content and thus shaping the space for disagreement’ (2003: 267)—the interpretation employed by a non-professional may be rather different from that arrived at by an analyst working within a particular understanding of argument. The ordinary arguer has concerns ‘organized around a sense of how to respond to another’s move and pursue controversy in a practical communication setting’ (2003: 268) unencumbered and uninformed by theoretical standards of argumentation. It is not necessarily a problem that the scholarly analyst is influenced by his own conception of argumentation, but it is an influence of which he should be aware. Aakhus’s remarks relate in 3 I leave out a fourth, ‘notation’ as being specific to the form of analysis with which he is concerned.

10.3 Reconstruction

159

particular to pragma-dialecticians whose concept of the critical discussion has great sway over how they interpret a piece of argumentative discourse; followers of other theories may be less affected, but it should always be kept in mind that one’s own preconceptions of how an argument should look may affect one’s interpretation of a particular argument text, and that others, with a different theoretical outlook, or none at all, might see things differently. One generally accepted foundation stone of argument reconstruction is the principle of charity, whereby texts are given the most generous interpretation, so that reconstructions are as strong as possible. This is an idea found in Davidson (1984) and his theory of radical interpretation, discussed in Chapter 1. The well-known formulation of the principle as applied to argumentation theory comes from Michael Scriven: ‘The Principle of Charity requires that we try to make the best, rather than the worst, possible interpretation of the material we’re studying’ (Scriven 1976: 71). This means adjusting the actual text to make it into something more like what a reasonable arguer might have said. Without such a principle one may find that arguments which one has rejected can very easily be revived by making small adjustments to their phrasing, meaning that one saves time by assessing the strongest version one can create in the first place. This idea has been extensively discussed, in an exchange among Trudy Govier (1981, 1982), Jonathan Adler (1982) and Ralph Johnson (1981), as well as by others, and the conclusions of the debate around the passage are helpfully summed up by Marcin Lewi´nski thus: ‘(1) some kind of basic charity is necessary and unavoidable in argument analysis and evaluation; however (2) strong charity (as typified by Scriven’s formulation) is methodologically and practically unsustainable; therefore (3) some limited, modest, or moderate charity should instead be adopted’ (2012: 408–409). Lewi´nski goes on to raise plenty of doubts about the validity of the principle in the interpretation of argument, concluding that it creates a paradox, since: ‘An arguer is either the protagonist or the antagonist, and an analyst who grants charity to but one of them, easily becomes uncharitable to the other’ (2012: 435). This leads him to advocate only a ‘minimal’ application of the principle, and that with caution. There is a reflection here of the Relevance Theory of Wilson and Sperber, discussed in Chapter 3. On their approach, ‘the implicatures of an utterance are those contextual assumptions and implications which the hearer has to recover in order to satisfy himself that the speaker has observed the principle of relevance’ (Wilson and Sperber 1986: 250, my emphasis). We may say that arguing well, or at least coherently, is a special case of being relevant; therefore, when we search for the argument implicated by the utterances of another speaker, we too should make the assumptions which we have to, and only those we have to, in order to make those utterances relevant. In the procedure given below, there are no specific instructions for the reconstruction of arguments, but this minimalist application of a charitable approach is suggested by the fact that texts are reformulated in order that they may pass the next stage of analysis and no more. Whether the argument which is finally accepted, possibly after several reformulations, is actually the one that was made by the original proponent is a separate question. That argument may be regarded as the closest

160

10 Analysis of Arguments

possible form to what was actually said which does not fail at any point in the analysis. Any reconstruction that does take place should be justifiable; if it is not, an opponent may well state that the claim is accepted, but deny that it is, in fact, the claim which was made previously and with which they disagreed. The most difficult issue with reconstruction for an analyst paying particular attention to the linguistic features of the argument is the necessity to put new words into the mouth of the original protagonist. There is little that can be done to avoid this, although the following guidelines are intended to keep the impact of the reformulation to a minimum: 1. Where possible, use the same vocabulary and/or structures found in the original text. 2. Be sensitive to argumentativity and implication inherent in the reconstructed language. 3. Do not attempt to ‘improve’ the expression beyond that which is strictly necessary. 4. Be as precise as the original text allows. 5. Where possible, seek the agreement of the arguer to the reconstructed version. The whole reconstruction process may be summed up thus: On the analytical level, the central question is how argumentative discourse can be reconstructed in such a way that all those, and only those, aspects are highlighted that are relevant in view of the ideal model that is chosen as a theoretical starting point. Such a reconstruction aims at a calculated merger of the ideal and the real that satisfies both the normative requirements exemplified in the ideal model and the descriptive data of empirical reality. (van Eemeren and Houtlosser 2015b: 118)

The section below will briefly consider where those normative requirements come from, and how they can be justified.

10.4 Normativity Frans van Eemeren describes informal logic as ‘a normative approach to argumentation in everyday language that is broader than the formal logical approach. The informal logician’s objective is to develop norms, criteria and procedures for interpreting, evaluating and construing argumentation that are faithful to the complexities and uncertainties of everyday argumentation’ (2015: 12). That description fits very well with my purposes, and with those of this chapter in particular. Something more needs to be said, however, as to the origin of those norms. There are two in ways in which one might justify labelling some arguments good and others bad. The first would be to claim that a certain standard of reasonableness is intrinsically good and that claims which flowed from the application of this standard were based on good arguments and those which did not were not. This is obviously problematic from the philosophical point of view and more recent argumentation

10.4 Normativity

161

theorists have generally leant towards the second option, which is to define a goal for argumentation and then approve of that which helps achieve that goal and disapprove of that which does not. I shall look at two variations on this theme as being representative of the two broad approaches. Harvey Siegel and John Biro (1992, 1997) put forward a theory of ‘epistemic normativity’ in argumentation. Their conception of good argument is based upon the ability of premises to warrant belief in conclusions: ‘we cash out normativity in epistemic terms: arguments aim at the achievement of knowledge or at least of justified belief; good arguments warrant their conclusions, bad arguments fail to do so’ (1997: 278). Good arguments lead to knowledge, and fallacies are ‘epistemic failures’ which fail to do so. They also suggest that arguments are supposed to provide ‘good reasons’, which is beginning to look a little circular: a good argument is one which provides good reasons to believe its conclusion, or perhaps, good reasoning is that which employs good reasoning. This approach ultimately falls back on the essential value of rationality described in the previous paragraph: ‘fallacies– and failed arguments more generally – must be seen as failures of rationality. Hence the centrality of the normative notion of rationality to argument theory: an argument succeeds to the extent that it renders belief rational’ (1997: 278). Although they discuss various less important objections to their theory in the 1997 paper, it seems to me that the crucial flaw is this: if good arguments are to aim at knowledge, then a weak argument with a true conclusion is a good argument, unless we can claim that it doesn’t provide knowledge because it doesn’t provide a proper justification for the belief. This then leads to the question of what would be a proper justification. Presumably, a good argument. If, on the other hand, good arguments give cause to hold a belief rationally, then rationality becomes that which is supported by good argument. How then can arguments themselves be held to standards of rationality? This way of thinking is an extension of formal logic: a good argument in formal logic is one which is valid according to the a priori laws of logic, to be rational, in logic, is to follow those laws. In informal argumentation, however, there is no clear standard of what is rational, hence a good argument is a rational one, and a rational argument is a good one. The way out of this problem is to take a more rhetorical view of argument and assign a pragmatic goal to argumentative discourse: good arguments are those which help to achieve that goal. To a degree, this is the approach taken by pragmadialecticians, who often refer to the study of argument as ‘normative pragmatics’. Since the goal of argument discourse in the pragma-dialectical approach is the resolution of a dispute through a critical discussion, the rules for that type of discussion are aimed at encouraging argument moves which aid that resolution and outlawing those which obstruct it. Scholars who follow this tradition, however, are unwilling to accept that just any resolution is a good resolution, and cling to a standard of rationality. Siegel and Biro take issue with the statement from van Eemeren and Grootendorst that argumentative discourse should still be held to ‘a certain standard of reasonableness’ (van Eemeren and Grootendorst 1992a: 5) and their suggestion that argumentation be seen as ‘a rational means to convince a critical opponent and not as mere persuasion’ (1992a: 10). This, they claim, means that pragma-dialectics

162

10 Analysis of Arguments

is, in fact, embracing an epistemic view of reasonableness, and, if it is not, it is merely descriptive or rhetorical. Exactly how to understand the norms of reasonableness in the pragma-dialectical approach is not an easy task. Van Eemeren and Grootendorst state quite clearly: There is no need to assume the existence of an absolute and definitive form of reasonableness. Within the critical perspective, reasonableness is viewed as a gradual concept. The extent to which a particular rule is considered reasonable depends on the adequacy of that rule, as part of a procedure for conducting a critical discussion, for solving the problem at hand. (2004: 16)

The key point here to defend the pragma-dialecticians from the criticisms of being purely rhetorical in their normativity, is that although the criterion of reasonableness is usefulness in solving the problem, it only applies when solving that problem through a critical discussion; and that is viewed as ‘a discussion procedure in the form of an orderly arrangement of independent rules for rational discussants who want to act reasonably, the aim of formalization is reminiscent of the geometrical approach to reasonableness’ (2004: 16). There are, then, two different notions of rationality at work at different stages: when two disputants wish to resolve a dispute in a way which both consider to be reasonable, that is, which both consider to be based on rational behaviour and likely to deliver a fair and justified resolution, then they enter into a critical discussion. The discussion is structured in such a way that what is reasonable within it is that which will best lead to that resolution. Therefore, the whole institution of the critical discussion, standing for what it does, is somehow subject to a more logical, geometric, epistemic notion of rationality, but its internal workings are governed by a reasonableness characterised as suitability to achieving a discursive, pragmatic, rhetorical goal. These observations must now be applied to the procedure for argument assessment presented below. The first level, of Process, employs two sources of normativity. One is the explicit rules of the process being engaged in, for instance of parliamentary debate, where such rules exist. The second is unashamedly rhetorical and pragmatic: for the process to proceed, certain norms of conversation, such as relevance, must be adhered to. Reasonableness at this stage is synonymous with civilized, sociallyacceptable behaviour. The second level, Reasoning, which, like the first, is not fully expanded in this work, also relies on two principles. The evaluation of premises is based on truthvalue: an elaborate argument justifying the rejection of false premises is, I suppose, not necessary. The principle that a good argument is one in which the premises support the conclusion is inherent in our understanding of what an argument is, and that support cannot be present if the premises are not true. The same principle applies to the test of the strength of the inference—what is referred to below as evaluation of the lever—although it may appear at first sight to be more problematic here. When considering any argument, a judgement about the strength of the inferential relationship it expresses is paramount, and the necessity of making such a judgement is exactly what the lever analysis makes plain. What sorts of inference the analyst

10.4 Normativity

163

should accept, and how strong any inductive inferential force should be, are not matters which are touched upon by the process, however. The action of the lever should be rational, in the opinion of the person analysing it, in the context in which the analysis is being made. If the inference appears too weak, if it is not found to lead to a rational belief in the conclusion, then the analyst has a reason to reject the argument. That someone else may call upon him to justify that opinion is accepted, and some help in that justification is provided by the more detailed description of the evaluation process in Section 10.4 of the following chapter. The third level, that of language, which is the focus of this work, does impose standards, and does propose rejection of arguments on the basis of their not meeting those standards and, thus, being considered ‘bad’ arguments. The norms applied here are not those of rationality, but those of communication, and are, therefore, rather more easily defended. The sub-levels of analysis dealing with vagueness, ambiguity, and coherence all demand that the premises and conclusion, with no exception, are expressed in language which makes sense, which can be understood, and which does not allow of a variety of interpretations. That language should be comprehensible to its audience in order to be communicative is a requirement which does not call for an elaborate defence. The final layer in the analytical construct, the so-called ‘deep’ linguistic analysis demands that the nature of language, as described in the early chapters of this book, is understood and respected by those making arguments. It is a norm of any linguistic community that words mean what they mean and not something else which someone using them would like to them mean; and it is in the nature of language that it may play tricks on the unwary and cause them to reason awry, due to a mistaken grasp of that meaning. Naturally, the analyst who rejects an argument at this level of ‘concept’ analysis must accept a burden of proof in showing why the arguer is wrong in his conception of the word or phrase in question. Lastly, the examination of implicatures requires no normative basis as it is a mere unpacking of the meaning of a sentence and there is no rejection done as an immediate result. The tool for assessment described below is normative in that it gives justification for the discarding of certain arguments, but its key linguistic sections are based on norms of language use, what is necessary for successful communication, rather than norms of rationality.

10.5 Other Methods of Evaluation—A Brief Comparison While there is no space here to give a thorough description of how the evaluation of arguments might be carried out by analysts following other methodological approaches, I shall point to several differences which might have a material influence on both the procedure and the outcome of evaluations so made. The first, and most obvious, comparison to draw is with the process of a pragmadialectical evaluation. What will be seen is that although the overall procedure, described in the chapter below, follows a different conception of argumentation, the Informal Argument Semantics, which is the focus and principal contribution of

164

10 Analysis of Arguments

this book is entirely compatible with pragma-dialectics. To perform a basic pragmadialectical analysis of an argument dialogue the two principle operations would be to identify the four stages of the critical discussion and to categorise the speech acts committed therein. This allows the dialogue to be described in terms of the structure of the critical discussion and for the moves made by the participants to be analysed. The goal of this process, as an evaluation, is to compare the behaviour of the participants with the norms of argument laid down by the theory. One participant may then be found to have committed a fallacy by breaking those rules or both may be found to have behaved within the expected norms. This is, of course, a much simplified description, but it allows me to draw out key differences with the process described in the following chapters. The first is that the pragma-dialectical evaluation is, first and foremost, an evaluation of the behaviour of the protagonist and antagonist, not of one argument, understood as premise/conclusion set, abstracted from a discourse. The rules of pragma-dialectics, for arguers, then, cannot be set in opposition to the norms of the procedure below, for arguments. The two forms of assessment are not comparable in that they are not assessing the same thing. Indeed, rules 7 and 8 of pragma-dialectics, which refer to the reasoning of the arguments put forward, could be evaluated by means of the Reasoning evaluation process described below, and rule 10, relating to the use of language, could be enforced by means of the Language analysis, based on the Informal Argument Semantics. What is also clear from the brief description above, is that the process of pragmadialectical analysis is particularly well-suited to the study of dialogues, which is hardly surprising. The whole theory is constructed to allow the examination of interactions between those putting forward standpoints and those questioning them. That more traditional approaches to arguments ignored this dimension, and by so doing ignored the vast majority of our everyday experience of arguing, is undeniable. However, where the need to evaluate arguments ‘in the abstract’, so to speak, does arise, a procedure is necessary. Even this abstraction is not done without due awareness of the context in which the argument is made, it is though the argument which is being evaluated in my procedure, not the arguer or the dialogue. The examples which I use in Chapter 12 to illustrate the working of the procedure are ones in which a full dialectical situation is not present. It would be difficult to describe them as forming part of a critical discussion, and it is not clear who the antagonist would be standing in opposition to the standpoint. That is not to say that a pragma-dialectical evaluation could not be carried out, but it would be likely to focus on the same aspects as the procedure shown here, given that lack of dialectical interaction. There is, I think, at least some cause to agree with J. Anthony Blair that “dialogue is not an adequate model for all types of argument” (1998: 325). The comprehensive assessment procedure described below is applied to arguments—premise/conclusion sets—not to wider debates or discussions. A result of this is that counter-arguments are not considered: a counter would be considered on its own merits separately once the acceptability of the first argument has been established. This is a difference with other approaches. Wider argument discourse

10.5 Other Methods of Evaluation—A Brief Comparison

165

is often analysed through the use of diagrams to establish relations between arguments and complex series of arguments may be categorised as ‘linked’ or ‘convergent’ (Govier 2010), or, alternatively, as ‘multiple argumentation’, ‘coordinatively compound argumentation’, or ‘subordinatively compound argumentation’ (Snoeck Henkemans 1992: 15). For the latter author, these distinctions and the relationship between the arguments which make up the series are of vital importance as her book is, as the title shows, dedicated to Analysing Complex Arguments. The ambiguity of the word ‘argument’ begins to show itself again here. A complex argument is made up of arguments. It is only these ‘simple’ arguments which are evaluated in the procedure below; but, again, that does not put the procedure in opposition to approaches which examine complex arguments, it simply means that the relations between or among arguments are not considered here. There is no reason why an analyst who produces diagrams of wider argument structures cannot then evaluate individual elements of those structures in the way I describe. The work of Shiyang Yu and Frank Zenker on a ‘complete argument evaluation’ (2020) is a recent development which takes a more similar approach to what an argument is. Yu and Zenker attempt to provide a complete set of critical questions to argument schemes, which involves a detailed meta-level account of what schemes are, developed along the lines of Toulmin’s description of argument structure, leading to a categorisation of three types of attack on an argument: there are ‘three targets premises, inference, conclusion’ (2020: 19), although all types attack the conclusion directly or indirectly. This means that there are three types of critical question: CQ1— Are the data correct? CQ2—Is ‘If D, then C’ correct? CQ3—Is the claim correct? Although their critical questions are described as a ‘complete’ evaluation, this approach could be combined with my own framework and addressed to the analysis of the Reasoning aspect of argumentation. There is, however, one difference, reflecting that mentioned above concerning multiple arguments. Yu and Zenker have a sub-question: CQ3.2—Are there other arguments against the claim? This, it seems to me, is to bring in a distraction which prevents the completion of the argument evaluation. Since the reasoning analysis is only the middle section of my comprehensive procedure, it is not a question I would wish to ask at this stage: only once an argument has passed through all three stages of analysis would I consider other arguments which may be the basis for an attack on the conclusion. Other evaluative approaches set out broadly similar norms, but generally do not contain systematic procedures for their application. Trudy Govier puts forward her ‘ARG conditions’, where premises must be acceptable, relevant, and provide ‘good grounds for the conclusion’ (2010: 87). In the section describing the use of these conditions in evaluating arguments she writes: You first put the argument into a standard form so that you can see exactly what its premises and conclusion are. Then you explore whether its premises are acceptable. […] Ask yourself whether the premises are relevant to the conclusion. […] Ask yourself whether the premises, taken together, provide good and sufficient grounds for the conclusion. (2010: 94–95)

166

10 Analysis of Arguments

Which is good advice for novices, but does not constitute a systematic procedure for evaluation. The same might be said of other sets of norms such as the RSA— relevance, sufficiency, acceptability—of Johnson and Blair (2006). Both the ARG and the RSA are ways of describing the key elements of argument acceptability which lie behind the framework of my evaluation procedure, but they are not procedures in themselves, and the results of their application are not as clearly defined as those of the systematic process of evaluation given in the following chapter. Finally, it should be noted that none of the approaches to evaluation described here is concerned with a detailed linguistic analysis of arguments. Pragma-dialectics, though focussed on arguers more than arguments, mainly describes what I assess under the Process analysis, and the others mentioned deal with what I call Reasoning analysis. It is for this reason that this work places most emphasis on the language analysis, the Informal Argument Semantics, which constitutes the third level of argument analysis in the comprehensive procedure. Other approaches to evaluation may take a different view of the first two levels, but the third, I believe, can be a complement and an extension to all.

Chapter 11

CAPNA—The Comprehensive Assessment Procedure for Natural Argumentation

The principal aim of this book is to set out a way of fully analysing the semantic content of arguments. That, however, is certain to be a difficult and laborious process. I have stated more than once that the fallacies of expression are among the subtlest, and occur most insidiously at the highest level of argumentation. The vast majority of suspect arguments will not require this level of forensic examination, but, of course, a systematic method for deciding which do make it through to the final level of evaluation is required. The CAPNA is designed to fulfil that task. There are three points to make about the name which I have given to the procedure. I consider it to be comprehensive because it takes into account the three elements of argumentation which I earlier defined. This does not mean that it is complete and cannot be expanded upon or elaborated, rather that, as a procedure, it focuses on the entirety of the argument, not any one part. In this work, I have placed emphasis on the Informal Argument Semantics developed for the evaluation of language, but this is only one element of the whole. Secondly, I have used the terms assessment, analysis, and evaluation more or less interchangeably throughout, but should now note an important reason why the CAPNA is a procedure of assessment. The procedure assesses natural argumentation, but it is the person using it, the analyst, who makes the evaluation. The description of an argument as good or bad is a value judgement made by a person on the basis of the assessment done via the procedure. The precise semantics of the words here are of less importance than the distinction drawn: the procedure itself is intended to be systematic and lead to reproducible assessments, the evaluations made on those assessments will be more subjective. The assessment procedure is, therefore, a tool in the evaluation of the argument. Lastly, the term natural argumentation is short-hand for naturally-occurring, and is used to make it clear that the CAPNA is not constructed to deal with formal arguments or contrived examples, but is intended as a tool for use with ‘real’ texts. This use of natural argumentation should not be confused with the more specific sense in which the term is employed by Sally Jackson and Scott Jacobs, who distinguish it from naturalized or ‘built up’ argument (Jackson 2015; Jackson and Jacobs 2006) © Springer Nature Switzerland AG 2021 M. Hinton, Evaluating the Language of Argument, Argumentation Library 37, https://doi.org/10.1007/978-3-030-61694-6_11

167

168

11 CAPNA—The Comprehensive Assessment Procedure …

which occurs through processes of socialisation, but is also present in real-world conversations. Figure 11.1 gives a graphical overview of the whole procedure, and the four steps in the analysis are described in detail in the following sections. It should be stressed once more that the evaluation is of arguments, and arguments alone, not of moves in an argumentative discussion; and that acceptance of an argument does not entail ultimate acceptance of its conclusion, as stronger counter arguments may be found to exist. The CAPNA is not designed to weigh the relative strengths of conflicting arguments, although its use will almost certainly throw into clear relief the points at which any particular premise to conclusion inference exhibits weakness and is vulnerable to attack. Acceptance of natural argumentation is always defeasible, and rejection should be treated as a stimulus to encourage the better formulation of argumentation rather than as reason to dismiss the standpoint of which it was offered in favour.

11.1 Initial Analysis The scheme begins not with an argument, but an apparent argument. The first step is to decide if there actually is an argument in the subject text which can be assessed. The initial analysis is an important stage, and incorporates elements from all the other stages: an outrageous display of ambiguity such as in the examples one sometimes finds in textbooks of an equivocation on semantically unrelated homophones, would be enough to dismiss the text at once; likewise with flagrant abuses of process such as outright insults. More sophisticated texts might also be rejected on the grounds that they do not actually argue for anything, even though the language and circumstances lead one to expect that an argument should be contained within them. This stage can require a good deal of work: longer texts will often contain many apparent arguments, but they may not come together to form an overall argument. One would like to suggest that this is a feature of inexperienced writers, but the example discussed below of an article by The Guardian’s Simon Jenkins on British universities suggests that is not always the case. Initial analysis may discover that texts which claim, or are thought, to be arguments are not in fact arguments at all, since they advance no standpoint on the basis of any inference. It may also discover, and this will frequently be the case, that the text as it stands does not contain an argument, but could be restructured to contain one. This may be a result of a certain amount of muddled thinking, where the producer of the text appears to be trying to get a point across, but requires help in its formulation, or simply an utterance with an implied argument which is clear to any competent speaker of the language in which it is expressed, but which cannot be properly assessed until it is made explicit. That is why in Fig. 11.1 rejection at the initial stage, and indeed all stages, leads to possible reconstruction rather than outright rejection. Indeed, it may be the case that very few texts are accepted as complete arguments without any form of reformation taking place. Even once this has been done, further analysis

Fig. 11.1 CAPNA. A comprehensive assessment procedure for natural argumentation

11.1 Initial Analysis 169

170

11 CAPNA—The Comprehensive Assessment Procedure …

may require a second or third round of reformulation since the need to reconstruct may come about as a result of deficiencies at any of the levels: process, reasoning, or expression. The initial reconstruction of the apparent argument is aimed at making it an actual argument, not at making it a good one. The question may arise as to how an argument is to be identified. To some extent that is at the discretion of whoever is employing the procedure, and will depend on how the strength of the reasoning is going to be assessed. In what follows, the reasoning evaluation is made using principles taken from the Periodic Table of Arguments approach of Jean Wagemans; an approach which has a fully developed Argument Type Identification Procedure.1 This procedure establishes both that there is an argument, and what type of argument it is, making it an invaluable tool. The Wagemans approach is discussed more fully in Sect. 11.3 below. In a way, the initial analysis stage is a general assessment scheme within a general assessment scheme and the structure may be thought of as resembling a set of Russian dolls, with analysis being taken ever deeper and becoming ever more precise at each iteration. This may appear to be in danger of becoming a never-ending process in which there is always a larger doll to swallow up the smaller and no argument can ever escape into acceptance: each full analysis becomes the initial analysis for the next run through the scheme. This is not, in fact, problematic at all. The analysis is designed to find flaws in the argument, once it has done so, the process is halted. If no flaws are found, the doubter can continue to apply the procedure, but, having had a good look and tested it once against all the listed criteria, is obliged to accept the argument until such time as an error is discovered in it. This is the pattern employed in scientific theorising: experiments seek to disprove theories rather to prove them, and every theory which is accepted after some initial consideration is held true up until the moment an experiment can prove it false. So it is with arguments. If the reasoning is considered defeasible then this vulnerability to new evidence is expressly acknowledged, if it is categorised as deductive, it is still open to criticism in further analysis and understanding at the process and language stages: a better understanding of the situation, perhaps of the argument to which one is responding, may lead to a deductive argument’s being rejected as irrelevant, for instance. Once one careful analysis has been carried out and no errors found, the proponent of the argument has the right to expect his argument to be accepted, but a determined opponent, while being obliged to concede that he can find no flaw or fallacy with which to deny the truth of the conclusion, is at perfect liberty to continue considering that argument, in his spare time, as it were, and to return to it at a later date with ‘new evidence’: we are not condemned to accept for all time any expressed reasoning with which we cannot find fault at first sight, though the belief that further examination might reveal a flaw is not a fair justification for rejecting it either. It is important to remember that an argument’s being accepted does not mean it is the clinching point in a debate, it does not mean that it cannot be replied to, or outweighed; only that, in itself, it is acceptable and deserves to form a part of our consideration. 1 See

https://periodic-table-of-arguments.org/argument-type-identification-procedure/ for the latest version.

11.1 Initial Analysis

171

One final point to make about the initial stage of analysis is that it is, as with any kind of textual analysis, heavily dependent on the understanding of language of the recipient of the argument. This may seem at odds with the fact that the close analysis of language does not take place until the final stage of the process. What needs to be made clear is that this scheme is intended for use by competent speakers of the language of the text; it is not intended for computers and would be of little use to anyone who couldn’t comprehend the arguer’s words. The later stage of linguistic analysis is intended to discover subtle semantic factors which are not obviously apparent, even to the most competent speakers, upon a first hearing of the argument. This is of particular importance when an argument, or elements of it, are expressed ironically, or in some other way where the speaker meaning is not the same as the literal meaning. While part of the linguistic analysis stage of the scheme is to tease out significant implicatures of the premises and conclusion, that should only be necessary when those implicatures are not immediately apparent. When the ‘real’ meaning of an utterance is different from the literal meaning of the words used, in a conventional pragmatic way, and the main content being expressed is actually an implicature only of the words used, the assessor of the argument should see that immediately and recognise that reconstruction into a more straightforward form of expression is necessary. Ideally, that reconstruction should be carried out by the original proponent, but this is not a necessity. If the irony goes unnoticed, then the apparent argument may simply be rejected as containing no real argument, or may go forward in its original form, taken literally, and is likely to be rejected at a later stage. In either case, the proponent, if available, or any supporter of what was ‘really’ being said, can respond by pointing out the correct interpretation of the text and demanding a new analysis. When an apparent argument is misunderstood and the form of the utterance taken too literally, but there is none who might correct this misinterpretation, then an error of analysis has occurred: no procedure can save an assessor who doesn’t understand his own language, and no exchange of words can be completely free of the danger of misinterpretations. This applies also at a wider discourse level: it may be that certain parts of the argument are embedded deeper in the conversation, or are taken as given by everyone involved in the debate. Although it has been stressed that the CAPNA is not a tool of discourse analysis, the identification of arguments obviously requires that the analyst consider the whole context in which the apparent argument is made, in order to make a fair assessment. Much has been written about the moves of argument discourse, and there is no need to review it here as a guide to conducting an initial analysis, suffice to say, that this type of discourse awareness is not forgotten and is an essential part of the analyst’s skills.

172

11 CAPNA—The Comprehensive Assessment Procedure …

11.2 Process Analysis Once an apparently argumentative text is accepted through the initial stage, in its original version or after some reconstruction, it can be analysed in the context of the first of the three stages of argumentation: the process. I do not here provide a full breakdown of this stage. What exactly is examined will depend upon the particular process in operation, a set of evaluative procedural questions, therefore, needs to be developed for each individual mode of argument, taking into account the special rules and features of the situation under consideration. Certain elements, however, are of a more universal nature and some of these are captured in the rules of Pragmadialectics. Recalling that only utterances which are identified in the initial analysis as being, or containing, arguments are considered here, the generally applicable constraints would fall into two categories: relevance and appropriateness where they would be joined by the particular requirements of specific argument processes. A good deal was said about relevance in Chapter 7, little of it encouraging. It is clear that to be acceptable an argument should be relevant both logically and dialogically: that is to say, that it should address the topic at hand following a chain of inferences which can be connected by others, and also that it should comply with some form of Wilson and Sperber’s Principle of Relevance, discussed in Chapter 3, which is generalisable to all types of discourse. At this stage, however, that is as far as I shall take it. Similarly, appropriateness has two aspects: the argument should be appropriate to the business of the argument mode; and it should be appropriate socially, not breaking the rules, written or otherwise, of what it is permissible to say and how it is permissible to say it, within the given context. It might seem somewhat prudish to suggest that an argument containing obscenities, for instance, should be rejected just because it offends the sensibilities of some who hear it, and yet that is what would happen in a court of law or parliament. It is worth pointing out once more that while these concerns are reflected in the rules of pragma-dialectics, those rules relate to the behaviour of parties to the discussion, not to arguments themselves which those parties have put forward. The emphasis in this work on arguments and their semantics is not designed to play down or ignore the actions of parties to argumentation, but to provide a different perspective of analysis. A simple scheme for Process analysis is presented in Fig. 11.2. A full Informal Argument Pragmatics to match the Informal Argument Semantics presented in Sect. 4 below, remains an ambition for the future. As to the overall structure of the assessment procedure, it is, I suggest, logical that the text be examined first for procedural compliance, since any argument, be it weak or strong, well-expressed or muddled, is of no interest if it does not follow the rules of the procedure at hand and is not admissible to that process. If rules are broken, then the argument is rejected as having committed a process fallacy. Yet, again, it is important to stress that despite the layout of the procedure, its sections should not be viewed in isolation and its flow should not be restricted to an unchanging linear progression. The straw man fallacy was described above as often relying on a misinterpretation, deliberate or accidental, of the words of one’s opponent (it may

11.2 Process Analysis

173

Fig. 11.2 A scheme for the analysis of argument process

also rely on an outright lie for which the words used in expressing the position under attack provide no support on any interpretation). Since straw man argumentation fails through the fault of irrelevance, and only pretends to be a response to a standpoint already made, it would be considered a fallacy of process. How then does one identify this error without having yet reached the stage of linguistic analysis? This is an interesting question and takes us back to the consideration of the possibility of dividing argument from meta-argument. Straw man fallacies are always meta-arguments in the sense that they are always, by definition, responses to the arguments, or at least standpoints, of others. One way of dealing with this would be to say that in order to analyse any argument which is put forward as a riposte, one must first have an analysis, at least in the initial stage, of the original standpoint which it is countering. That initial stage analysis would make it clear what the actual argument was, and the claims being made would be compared at the process stage under a relevancy analysis, with the rebuttal being rejected if it did not rebut the actual claim first made. Alternatively, obvious examples of misrepresentation, particularly where the original claim is actually restated in an inaccurate way, might be detected at the process stage and rejected there, while subtler instances might be allowed to progress down to the full linguistic analysis. The crucial element here would be for the entire argument—including the original argument being attacked—to be considered. In this way, differences in the uses of certain words would be exposed as equivocations, and differing implicatures of original and restated arguments would also come to light, making the shift in meaning between the proponent’s actual claim and the antagonist’s interpretation of it transparent.

174

11 CAPNA—The Comprehensive Assessment Procedure …

In any case, a Straw Man, or any other irrelevant argument, might be detected at the Process stage, or might advance and be detected at either of the remaining stages— its irrelevance might only become apparent when the reasoning or expression were examined more closely. This is an illustration of why the use of traditional fallacy names, while sometimes useful, is not part of the procedural design: an argument is fallacious at the point at which it is found to be fallacious—analogies between different instances of failures at the same point or between the overall structure of failures at different points may be of interest, but are not the concern of this approach to evaluation.

11.3 Reasoning Analysis Arguments which are accepted as rule-following in the given context are then subjected to a reasoning analysis. There are a number ways in which this analysis could be carried out. One way would be to try to match the structure of the argument with one of the schemes listed in Walton et al. (2008), or in a different such collection, although there is, of course, great disagreement amongst scholars on how many argument types actually exist. Once the argument type is identified, the associated critical questions could be asked to establish the perceived strength of the inference. Some arguments might be dismissed as clear reasoning fallacies, some accepted as well-formed and supported, and others might be passed on to the linguistic analysis stage with a caveat of possessing dubious strength. This whole process could be proceeded by a check for formal, logical fallacies, although obvious examples would likely have been noticed at the initial stage. One problem with this method is that such a classification does not rest on a theoretical but on an empirical starting point. The argument schemes that are distinguished in this approach have been observed in argumentative reality and have subsequently been subsumed under one of the three main categories. No theoretical rationale is given for the number nor the name of the categories in this classification. (Wagemans 2016: 2)

This means that while lists of schemes are useful for familiarising ourselves with the types of arguments we may encounter in actual discourse, and the types of responses we may give to them, they are a rather blunt tool when it comes to exact identification of argument type. A second possibility, as suggested above, would be to employ the categorisation of arguments suggested by Jean Wagemans, which has led to the development of his Periodic Table of Arguments (Wagemans 2016, 2019). The categorisation relies on three distinctions, the first of which is into two types: ‘subject arguments’ and ‘predicate arguments’. These may then be classed as either ‘first-order’ or ‘secondorder’, and finally the premise and conclusion are classified by type as statements of fact, value, or policy. An argument can, therefore, receive a classification such as 1 pre FF, if it is a first-order predicate argument and both premise and conclusion are

11.3 Reasoning Analysis

175

statements of fact (this is, in fact, the code for arguments from sign, from correlation, from cause, and from effect). This method of categorisation is considerably more systematic than those which have been attempted previously, and the placing of the argument types into a table with four quadrants allows the relationships between them to be seen at a glance, and also illustrates certain gaps: classifications for which argument types have not yet been found. There are a number of points of interest in Wagemans’s typology. Firstly, the arguments he works with consist of one premise and a conclusion, which goes against the strictly logical reconstruction of arguments usually employed and featuring at least two premises, as the warrant or major premise is left out. This omission, however, makes the arguments considered more similar to real world argumentation found in actual discourse, whilst not preventing the formulation of what is called the ‘lever’ of the argument, which is an expression of the support relationship between the premise and the conclusion of the argument. Unlike in traditional approaches, the lever is derived from the form and content of the actual discourse rather than added by the analyst to match an a priori determined logical form. One of the examples he has used as a typical first-order predicate argument in the papers cited will illustrate: The suspect was driving fast, because he left a long trace of rubber on the road. Which could be reconstructed: Premise: The suspect left a long trace of rubber on the road. Lever: Leaving a long trace of rubber on the road is an effect of driving fast. Conclusion: The suspect was driving fast. Obviously, the single sentence version is closer to typical real discourse, and the characterisation of the argument is carried out on the basis of what was actually said, not by imagining warrants which are necessary to complete the logic. This approach works because the table is one of classification not evaluation, while the formulation of the lever prepares the ground for an evaluation of the relationship between the two parts of the statement. The difference here is that the lever is a linguistic expression of a pragmatic feature of the argument that typically remains implicit in the discourse. It exists in abstract form for each type of argument, and comes into play only after the argument has been identified, at which point the abstract can be made concrete to fit the given example (Wagemans 2019: 64–65). Secondly, it is a feature of the linguistic structure of the argument which is used to make the initial division into subject and predicate arguments: when the subject is repeated, as in the example above (the suspect, he (the suspect)), then we have a predicate argument. This can be represented as ‘a is X, because a is Y’ (Wagemans 2019: 61). The first stage in understanding the reasoning used in an argument is not just a reading of the language of that argument, but an analysis of its structure. Second-order arguments are those in which there is no common term present in both premise and conclusion. The problem of classifying them as subject or predicate arguments is solved by the distinction made in linguistics between propositions and assertions: an assertion is a statement which carries with it a commitment to its truth on the part of the speaker. This means that the argument doesn’t contain just a proposition, for instance ‘a is X’, it contains the assertion ‘a is X is T’, where T stands

176

11 CAPNA—The Comprehensive Assessment Procedure …

for ‘true’. This allows arguments from authority, among others, to be classified as second-order, as shown in Wagemans’s example: For it can then be analysed as ‘We (a) only use 10% of our brain (X) (q) is true (T), because that (q) was said by Einstein (Z)’, which instantiates the form ‘a is X is T, because a is X is Z’ or ‘q is T, because q is Z’ and thus has the subject term (q) of the statements as the common term. (Wagemans 2019: 63)

Once the common term has been established in this way, second-order arguments examined at the level of assertion can be divided into subject and predicate types in the same way as those analysed at the propositional level. The third level of analysis defines each statement as being of fact, of value or of policy. These types are distinguished largely by the contents of their predicates, although the subjects of policy statements will generally be acts (A) rather than entities (E), meaning persons or objects. The typical form of the predicates can be characterised as: Fact—E has property P; Value—E is judged as J; Policy—A should be carried out (Wagemans 2016: 7). Clearly, such a classification also depends on a linguistic analysis of the predicates as being descriptive, evaluative, or normative. There are two advantages to this approach: firstly there is a more systematic way of assigning argument types to particular texts by looking at the three distinctions, subject/predicate, first/second order, value/policy/fact. In the Waltonian argument schemes there is some crossover and duplication, with some question as to whether they all really qualify as distinct argument forms, which is less of a problem on Wagemans’s typology. Also, to a large extent, the more traditional identification of an argument as a representative of a particular scheme takes place via an (meta-)argument from analogy, in which the text in question is found to be similar to the archetype. This is a process fraught with vulnerabilities over subjectivity and, indeed, possible doubts over the strength and reliability of arguments from analogy. Secondly, the three distinctions which are made in the Wagemans classification are of relevance in the following, linguistic, stage of analysis. A predicate argument, for example, is one in which the subject is repeated: this makes that subject an obvious place to look for a possible equivocation. The knowledge that an argument is one of fact or one of value may also affect how the language is assessed, given that the background assumptions implied in evaluative statements would be expected to differ from those supporting factual claims. In its basic formulation, the Wagemans tabulation does not contain an evaluative element: arguments are placed in the table and the positions are named, but there is no system for labelling which forms are suspect and which acceptable, or which forms are strong and which weak. However, a simple, yet systematic method for evaluating the underlying reasoning of arguments within the PTA framework has recently been developed (Hinton and Wagemans, forthcoming). In this method, an argument is identified and labelled, then subjected to procedural questions, specific to that form of argument, which evaluate the acceptability of the premise and the strength of the lever. A simple diagram illustrating this process is shown in Fig. 11.3. One further consideration in the analysis of the reasoning behind an argument is just how weak it needs to be before it is rejected as a reasoning fallacy. The

11.3 Reasoning Analysis

177

Fig. 11.3 A scheme for the analysis of argument reasoning

question of argument strength, which is addressed via both formal and informal approaches, is beyond the scope of this work, and there is no space here for a full review of the literature on the subject.2 Still, one or two comments on the topic are in order. Firstly, the scheme as it stands does not make any judgement on how best to evaluate the strength of an inference, leaving that decision to anyone who wants to employ it. It is equally amenable to formal calculations and informal assessments of argument strength. It could be expanded differently from the suggestion above, to first identify types of inference; deductive, inductive, abductive, for example, and then introduce criteria of acceptability for each type, but that would be a major task, unconnected with the core theme of this volume. Secondly, since the scheme is formulated in such a way as to make a binary accept-and-analyse-further/rejectas-fallacious distinction, it appears to be ill-equipped to deal with weaker forms of inference which are neither clearly acceptable nor downright fallacies. This weakness is only in appearance, however, because there is no pressure for any argument to be rejected at any stage: each level of analysis is an opportunity to reject, and the user of the procedure as a tool can decide how generous and forgiving he wishes to be. A question mark may be placed by an argument at any stage, even though it be allowed through to further analysis. If that analysis discovers a flaw requiring immediate rejection, then the earlier uncertainty is immaterial; if no such flaw is found, the original doubt may be revisited and resolved either through a qualified acceptance or a carefully justified rejection. Such qualification may also take the form of a slight reconstruction: Toulmin, remember, has qualifiers as a separate and distinct part of argument formulation. 2 For

the increasingly popular Bayesian view, see the papers in Zenker (2013).

178

11 CAPNA—The Comprehensive Assessment Procedure …

The two boxes ‘weak inference’ and ‘strong inference’ in Fig. 11.1, then, are open to interpretation and are relative to context and the process of which they are a part. Even within legal argumentation, there are different levels of proof for criminal cases—beyond reasonable doubt—and for civil cases ‘preponderance of evidence’.3 Clearly, what is strong enough in one context is too weak in another: on such matters the scheme makes no comment.

11.4 Informal Argument Semantics The final stage in the assessment is the linguistic analysis. This is the true destination towards which all of the preceding pages have been heading, and it will be described in detail in this section. Before introducing that procedure and the concept of Informal Argument Semantics which underpins it, it is worth noting how this analysis fits into the general scheme. The language analysis has three possible outcomes: the argument may be rejected, it may be accepted, or it may be reconsidered as a new argument once all the argumentative content inherent within the language used has been unpacked. In this case, the newly-elaborated argument is sent back to the first level of analysis in order to be assessed at all levels once again. It can be imagined that in a truly rigorous examination, this process would be almost inevitable. Rejected arguments are also considered for reformulation and reassessment before being conclusively discarded. Whether or not this can be done and whether it would be considered worthwhile depends on the exact nature of the linguistic problem in the argument and the situation. For example, the exposure of an ambiguity might cause the proponent of an argument to withdraw his standpoint, making further analysis unnecessary, but, in other circumstances, the new, disambiguated argument, might then go on to be tested to discover whether the ambiguity was essential to its reasoning or not: the mere fact of a word’s being ambiguous does not invalidate all reasoning processes in which it has been used. Informal Argument Semantics (IAS) is the name I give to the framework behind the entire process of linguistic assessment of argumentation. In some sense, the whole CAPNA is a semantics, as the word is used in formal approaches, of which more below; and also, in some sense, the Process analysis described above is ‘linguistic’, since it is a form of discourse analysis; the degree to which the process-focussed Pragma-dialectical approach is a theory informed by principles of linguistics has been discussed at length in earlier chapters. That is a linguistics, concerned with a discourse situation more than the words used, which is the realm of pragmatics. Here, in this final stage, we are concerned with semantics: with the meaning of words, with the meanings of words, and with the meanings of combinations of words. There are such things as formal semantics of argumentation. These tend to be based on Dung (1995) argumentation frameworks, which allow arguments to be depicted 3 How

these are understood in practice is another matter. See Shapiro (1991) and Demougin and Fluet (2006).

11.4 Informal Argument Semantics

179

graphically. ‘Given such a graph, one can then examine the question on which set(s) of arguments can be accepted: answering this question corresponds to defining an argumentation semantics’ (Baroni et al. 2011: 365). I cite this simple description from a review of various argumentation semantics, as it sets out clearly what is meant by a semantics in this case: a semantics is a tool for deciding which arguments are to be accepted and which rejected; or what the same authors call ‘a formal method to identify conflict outcomes for any argumentation framework’ (2011: 366). There are also a number of competing semantic systems for ranking arguments, thus overcoming the difficulty caused by the simple accept/reject judgement (Bonzon et al. 2016). The two papers here cited, while interesting and informative, are both very technical and neither contains even one actual argument: the gulf between formal and informal approaches to studying argument is wide and it isn’t easy to see how it might be bridged. Informal Argument Semantics, then, is not to be understood as simply an unformalised version of formal argument semantics. It is rather meant as a semantic analysis of informal argument: a way of assessing arguments based on a precise and detailed examination of the meanings of the words and structures used in the expression of them. That is not to say that no elements of formalisation could possibly be compatible with this framework. The relatively simple form of description logic developed by Patrycja Kup´s (2019) for identifying ambiguities in an argument is a variety of formal approach which is applicable to the analysis of a natural language text. It is, therefore, a formalised tool for assessing informal arguments and could be integrated into the ambiguity analysis section of the scheme given below, although my intention here is to avoid any use of formalisation: these schemes for assessment are designed to capture and reflect the process of argument analysis which already goes on in the mind of an experienced arguer, rather than provide wholly new ways to ‘compute’ the acceptability of expressed reasoning. The schematisation allows for greater systematicity and clarity in the exposition of that analysis, and exposes certain elements of it of which less experienced individuals may not be aware. As the examples discussed below will show, an argument often doesn’t sound quite right, but it can be very difficult at times to say exactly why. If one has knowledge of common fallacies, one may be able to identify the mistake, expose and explain it; but even then, as I have repeated throughout, fallacies of expression tend towards the subtle and are not given away by structural elements in the statement of the argument as reasoning fallacies often are. The assessment procedure allows the recipient of an argument to illustrate exactly why he chooses to reject it, or gives more confidence in accepting it, or, perhaps most importantly and most likely, provides a channel into which discussion can flow by exposing where the uncertainty lies and encouraging the proponent to engage with that doubt and attempt to guide his own argument through to the green light at the end of the decision tree. The scheme for linguistic analysis examines four factors in the semantic content of the argument. These are four different ways of approaching meaning, and ask four different questions. Firstly, there is clarity of meaning: is the meaning sufficiently clear and precise? Then, consistency of meaning: is there an equivocation at play? Thirdly, coherence of meaning: do the elements of linguistic structure fit together

180

11 CAPNA—The Comprehensive Assessment Procedure …

felicitously? Lastly, there is a deeper analysis: what is meant by the language of the argument beyond the surface meaning of the words used; what is implied; what is assumed; what has been hidden? This refers to the completeness of an argument. The last category also contains those errors which are brought about by a mistaking of the very nature of language—the matter at the heart of this work—when, for example, abstract nouns which are merely abstractions become confused with concrete nouns referring to something substantive, or cognates are assumed to be linked in use due to a common etymology. These issues may be put under the heading of conceptualisation. A simple way to lay out the principles of language use in arguments which the four stages of analysis are designed to examine, then, is to list the five Cs of informal argument semantics: Clarity—language must not be vague. It must be clear enough for the purposes of the argument. Consistency—meanings must not alter within the argument. This prevents equivocation. Coherence—words may not be combined if they are not semantically compatible. Completeness—arguments must be given in full. This involves making explicit implicatures and unstated premises. Conceptualisation—the use of concepts within the argument must not be based on an erroneous understanding of the relationship between language and reality. These principles are captured in the four analysis categories, which, naturally, are based on the discussion of linguistic fallacies in Chapters 8 and 9. Some of the errors in argumentation mentioned therein as being related to language use, such as straw man fallacies or obviously loaded questions, might well have been discovered at earlier stages, but can also be detected within these categories through the inconsistencies and implicatures which come to light. The four stages listed here may not be exhaustive, although I interpret them widely and believe they capture the important areas in which linguistic error influences argument. This will become clearer as the internal structure of each level of analysis is revealed below. The addition of extra levels either alongside or within the four categories shown in Fig. 11.4 would not in any way disturb the foundations of the system, and an increasingly fine-grained analysis would be an inevitable result of the use of the scheme in particular areas where particular types of linguistic fallacy are found to be prevalent. The first element in the linguistic stage is clarity analysis. The discussion of vagueness in Chapter 8 described different ways in which a text can be vague. If the overall content is too vague for an argument to be identified, or simply makes no sense at all, then the initial analysis will lead to a rejection. This leaves the two types of unclear language described earlier as fallacious: language which is, intentionally or otherwise, obscure, oblique and obfuscatory; and language which is underdetermined, making it impossible to know what precisely is meant by the proponent of the argument. This second type provides another example of how interrelated the stages of the analysis are, since the point at which such vagueness becomes fallacious is dependent on the process of which the argument is a part. In a general discussion on moral

Fig. 11.4 IAS. An Informal Argument Semantics

11.4 Informal Argument Semantics 181

182

11 CAPNA—The Comprehensive Assessment Procedure …

behaviour one might be able to state that one should always try to help others. The full argument would look something like this: P1: It is morally right to help others. P2: One should do that which is morally right. Conclusion: One should help others. As a general precept that might be acceptable, but if the debate were around government policy, all the elements here would be far too vague to allow the argument to pass. What would it mean to help others in a political context? In what sense is government bound by the ‘morally right’? Vague arguments can certainly be sound, but they are bad arguments nonetheless, because they are of no use. The question to be answered when assessing such claims is what degree of precision is required in the case at hand, and that is a reason why the process of which the argument is part must always be kept in mind throughout its further assessment: it is not forgotten about after the process stage of analysis. Example 4 in the following chapter shows how the apparently unproblematic phrase ‘clear and obvious’ turns out to be insufficiently precise for the process in which it is employed. One might also think that part of the problem with the short argument above is an ambiguity, and that, by extension, imprecise terms are ambiguous. This would be to confuse two different phenomena, however. Ambiguous terms have more than one meaning and both meanings may be perfectly clear and precise. Such words only become a problem when there is no way to tell which of the two is meant, or, most importantly in argument analysis, when the same term is used within one argument with a different meaning. Vague terms cannot be guilty of equivocations because it is never clear, in any instance, what precisely is meant, so the second instance cannot be said to have a different meaning. In the example above, we cannot say that there is an equivocation on the repeated term ‘morally right’: we may claim that it is vague, even meaningless in the context, and one may certainly suggest that premise 2 is tautologous, making the whole argument circular, but there is no sense in which the meaning can be identified as one thing in the first premise and another thing in the second. This type of vagueness will frequently lead to outright rejection, or at least a very qualified acceptance, because it is difficult to remove through reformulation, since the reformed argument would need to show greater precision than the original; something only possible if the proponent has left other contextually relevant information with which their argument can be made more exact. In the absence of such information, any act of precisification can only be guesswork and while the subsequently formed argument may be offered up for re-analysis out of interest to the analyst, it cannot be ascribed to the original speaker. This difficulty is illustrated in example 1 in Chapter 12, where the meaning of the word ‘need’ requires some kind of clarification. The vagueness of poor prose is far more amenable to reformulation precisely because it is the original formulation which is at fault. Sometimes this is a very simple task: a pretentious and tiresome phrase such as ‘employment opportunity’

11.4 Informal Argument Semantics

183

which is in common, if completely superfluous, use, can easily be translated as ‘job’. At other times, this language will have been made deliberately obscure in order to hide its real meaning: apparently, the job title ‘Beverage Dissemination Officer’ has genuinely been used to describe the position of a barman (Wilkinson 2019). In such cases, one may or may not be able to guess the real meaning: where one cannot, and the phrase is significant to the argument, a rejection will follow on the grounds of a fallacy of vagueness. Whether or not such a fallacy has been committed is, then, partly dependent on the ability of the receiver of the argument to understand what has been put before him. The proponent of the argument may appeal against rejection, stating that it is perfectly clear to him what was meant. Such an appeal, however, cuts no ice. The responsibility of the arguer is to make claims in such a way that they are comprehensible to the audience: if they are not, then why make them in the first place? For this to work fairly, whoever is assessing the argument must be honest as to whether the language is really impossible to decipher or only aggravating in its opacity, and not continually call for greater clarification, which could be an effective tactic in delaying acceptance of an argument for which no other reason has been found to reject. Again, the procedure cannot guard against the prejudice or bias of the analyst employing it. Where reconstruction is possible, because the language lacks clarity, but the intended meaning can still be garnered from it, the new formation goes for reassessment. It might be suggested that this is unnecessary since the meaning was understood all along, but such language, especially when it has been deliberately constructed to side-step normal usage and present ideas in a light in which they are not usually seen, may have allowed the argument to pass through earlier stages by virtue of its obscurity. Imagine that the data premise in an argument is ‘I’m offering you a job’ or ‘I’m offering you an employment opportunity’. These two can be understood as having the same meaning, so no rephrase appears to be necessary even if the second form is needlessly verbose. However, the warrant premise could then be either ‘You should take every job that is offered in life’ or ‘You should take every opportunity that is offered in life’. An argument based on the first version would certainly be rejected as unsound, as clearly some jobs are best turned down; but an argument based on the second one might be accepted, given the positive way the word ‘opportunity’ is viewed. That, of course, is the whole point of using such language: to make the thing being discussed sound more attractive than it really is. This type of argument would be unlikely to make it through the ambiguity analysis in any case, as ‘opportunity’ is being manipulated (while we can use opportunity to mean anything which is on offer, that reading of the word is not the one suggested by the second premise), but the example is designed merely to show that earlier stages of the analysis can be affected by insidious phrasing, even where the genuine meaning is immediately apparent. Where no fallacy of vagueness is detected, the argument moves on to ambiguity analysis. Within this level is also included amphiboly, despite the fact that the scheme itself shows that rejection at this stage is due to an equivocation fallacy, a simplification for the sake of neatness. In truth, amphiboly is an awkward phenomenon to

184

11 CAPNA—The Comprehensive Assessment Procedure …

categorise and although it is different from vagueness, as described above, in that it leads to multiple possible meanings rather than an uncertainty of any meaning, amphibolous sentences could equally well be rejected on the grounds of containing a vagueness fallacy, as, ultimately, it is the lack of clarity in their construction which makes them unacceptable. As I suggested in the earlier discussion of fallacies, amphiboly is more likely to feature in argumentation as the subject of a dispute than a weakness in an argument, since usually one of the readings of the sentence will contribute to the force of the argument being made and the other will not. A simple example of what I mean: P1: A man-eating fish is dangerous (A man, eating fish, is dangerous). P2: A piranha is a man-eating fish (A piranha is a man, eating fish). Conclusion: A piranha is dangerous. If one were to hear this argument and question why a man should be dangerous when he’s eating fish, and how a piranha can be a man, one would not be taken seriously: it is obvious that the hyphenated version is what is meant, even when the argument is delivered aloud. The point being that an argument will generally ‘work’ when one interpretation of the sentence is used and fail with the other, making the intention clear. Still, a check for amphiboly is included. Ambiguity is a far more important issue, and an exceedingly complex one. As earlier chapters made clear, language is ambiguity. There can be no communication through signs unless those signs are capable of conveying multiple meanings. Every argument, therefore, will contain words exhibiting ambiguity and the role of the assessment procedure is to help establish whether or not that ambiguity is significant to the structure of the inference being made. Even then, that ambiguity is not of itself a reason to reject the argument since it may be that the proponent intends to include more than one of the possible meanings of a word within its scope. Take the simple, everyday English word ‘child’. There may be more, but we can state for now that this word has three meanings, all part of the same concept, not merely unrelated or homonyms or even far-removed uses of a once unified concept: a very small person, a person under the age of 18, the offspring of another person. Thus, a teenager may claim ‘I’m not a child!’ whilst legally still being one, and my parents still believe they have children deep into their retirement, even though those children became adults many years ago. There is clear scope for equivocation with this word: a 30-something who demands pocket-money from his mother and father on the basis of the principle that parents are responsible for the maintenance of their children is exploiting one. It is equally possible to state, however, that it’s good to have children because a child is a parent’s greatest joy, and to do so with an intended multiple meaning that child-birth, child-rearing, and the later taking pride in the achievements of an adult child are all parts of the greatest joy we can experience. The ambiguity does not have to destroy the argument, although, of course, it can. A further complication is that for an argument to contain an equivocation its original statement does not have to feature the same word or phrase more than once. Frequently, the word will only be repeated as a result of the reconstruction necessary

11.4 Informal Argument Semantics

185

to make the text into a full argument, as in example 3 below, where a good deal of work is necessary to produce an argument which flounders on the equivocal use of ‘education’. In such cases the repeated use of the word is only implied by something else which has been said. It is perfectly valid to reject arguments where the fully expressed form requires an equivocation, but there is perhaps a subtle difference between such cases and others where the proponent of the argument can actually be accused of having used ambiguous language. This type of ambiguity, where a concept seems to be understood in two ways, rather than an actual word, may be particularly hard for the arguer himself to notice. Another possibility is what I refer to as synonymic equivocation, which is exemplified in example 2 in the following chapter. There is no dual occurrence of one word with different meanings, but rather two words are taken to be ways of referring to the same thing, when, in fact, in the sense relevant to the argument, they do not. It is an accepted aspect of good style that words are not repeated where an alternative may be used instead, and yet, the full meaning of two synonyms is rarely, if ever, identical. For all the complexity of the phenomenon itself, the appraisal of ambiguity can be carried out in a straightforward and systematic way. The full argument, premises and conclusion, which has been earlier accepted as cogent and sufficiently precise, is now checked for repeated words or synonyms upon which an equivocation may rest. If it has been examined and placed in the Periodic Table of Arguments, certain terms will already have been identified as appearing in both premises. The question now is simply one of judgement as to whether the meaning of the word is the same in every instance. If the repeated term is a proper noun, there is unlikely to be a problem, when it is an abstract noun, it needs to be treated with some suspicion. This judgement rests entirely on the linguistic competence of the analyst. There is no criterion of meaning outside the understanding of the language user: a dictionary is of no use, it is for the analyst to decide if the word is meant in the same sense at every appearance or not. Even where no equivocation is found, or there is no repeated word to equivocate on, there is still a possibility of rejection due to ambiguity. Consider the following example said whilst looking at a photograph: P1: The man on the left is a traitor P2: Traitors must be shot Conclusion: The man on the left must be shot. There is no reason to think this is an equivocation: clearly the proponent of the claim means the same man each time he refers to the man on the left. However, the phrase is ambiguous—the left of the people in the picture or the left of the viewer?4 The argument can be rejected as a piece of practical reasoning since its final claim is ambiguous, even though in its structure it is a valid form. 4 This is precisely the confusion which led Barry Joule to mistakenly destroy a Francis Bacon picture

described to him by the artist as ‘the one on the left’ (Joule 2015). Had he been more alert to the possibilities of ambiguity, that work would have been worth many millions of pounds today.

186

11 CAPNA—The Comprehensive Assessment Procedure …

The third level of linguistic analysis is a simpler matter and one which is not often considered by other authors on linguistic fallacies. This is the question of linguistic compatibility, as exemplified in Chapter 8 by the claim that not reaching a deal could be rejected, when the item ‘not reaching a deal’ is not the kind of thing which can be rejected. If such constructions appear in the premises, the argument might be dismissed at the initial stage as being simple nonsense or at the clarity stage as being impossible of proper interpretation. If, however, the incompatibility occurs in the conclusion, and the reasoning has been accepted, then it has had a certain authority conferred upon it by that acceptance. It is surprising to consider that a claim which follows validly from two meaningful premises can itself be meaningless; and yet such is the case. When this does happen, then the argument, however good it may seem, has to be rejected on the grounds that its conclusion is not a comprehensible sentence. It is tempting to say that if that is the case, then there must be something wrong with the argument that has caused that incoherence; it cannot be just the incoherence itself, it must have come from somewhere. Sometimes, that may prove to be true, and a closer look may expose an earlier flaw, but it is in the nature of language that there does not have to be one. Natural language is not logical, two logical sentences combined logically can still lead to an illogical conclusion because of the way that words work. Just as vagueness and ambiguity are linked as both occur when meaning is not pinned down, so questions of compatibility are linked to the philosophical fallacies of language described below. The mechanism through which this incompatibility occurs is, just like those errors called philosophical, a product of the arguer’s having been deceived by the nature of language. There is, however, a reason for keeping them separate: the error should be visible to any competent speaker of the language in which it is made, its exposure requires only an understanding of the grammar of the words involved, rather than what I describe as ‘deep analysis’ in the paragraphs below. How deep is ‘deep’ is a legitimate question, as any argument which has got this far towards acceptance has already received some careful analysis. The language level itself is deepish, so to speak, and those not comfortable with delving into the terminology of grammar, may find themselves already drifting out of their depth. In these cases the likely cause of inconsistency is what might be named a ‘parts of speech generalisation’. This is a mistake where it is assumed that words may fit into slots in sentences purely because they are the correct part of speech, without realising that for many verbs, for example, the only thing the word has in common with other verbs is that it is classed in the grammar as a verb. Take the verbs ‘to eat’, ‘to die’, and ‘to believe’: they are in the same grammatical category, but if we can step outside of our common practice where we are used to treating them as being of one type in our speech, we can see that there is really nothing much linking them other than that they have a subject, different tenses, etc. Of course, as competent speakers of our language, we know that these words cannot all be used in exactly the same way: ‘to die’, in the meaning of experience death, is intransitive, it cannot take an object, and anyone who knows English will immediately reject a sentence which states that ‘I died him’ or some such ungrammaticality.

11.4 Informal Argument Semantics

187

Other examples are subtler, however. To return again to the ‘no deal’ rejection: the verb reject has a limited range of possible nominal phrase objects, and any argument that concluded that one should reject something which is not under consideration, such as ‘left-handedness’ or ‘clouds’, or not nominal such as ‘interesting’ or ‘at the top of the hill’, would be dismissed as nonsense at once. The case is made complicated by the fact that ‘a deal’ is very much the sort of thing one might reject, so it seems logical that ‘no deal’ ought also to be that sort of thing; but it isn’t.5,6 An example from a recent conversation illustrates the perhaps more common confusion which can be caused by treating verbs as things alike to one another. A colleague discussing motivations to act moved from motivation to do certain physical things, to motivation to believe a proposition. This led to a discussion over whether believing was the sort of thing one did as a result of being motivated. Is believing an act, in any sense like going to work is? I don’t see that it is, and so if ‘motivation to believe’ does have a meaning, I cannot see it: it is certainly not analogous to ‘motivation to act’. These levels, clarity, ambiguity, and coherence, allow us to say that a structure is provisionally accepted, and that accepted linguistic form can then be subjected to a deeper analysis in search of what I call philosophical reasons to reject the argument they express. The scheme for deep linguistic analysis set out in Fig. 11.5 is comprised of three levels. At the first stage improper definitional moves are investigated. Since definitions are usually signalled clearly, by ‘something is something else’ statements, as is the case in example 1 below, they are relatively easy to identify. The question then becomes one of judging whether or not the definition is legitimate in the circumstances. There is nothing inherently wrong with providing a definition although it may become the source of an equivocation where a redefinition and an original meaning of a word are employed within the same argument. This shouldn’t be too hard to avoid, since the whole point of providing a definition is usually because one wants to say precisely how one expects that word to be understood in the argument that follows, and yet, at the same time, one usually wishes to make an argument about something with which people are already familiar, not an entirely new phenomenon. Thus, when a scholar gives a definition of argument, it may be qualified carefully as simply his definition for his purposes—I have done this myself within these pages— but the intention is generally to say something about ‘argument’ and it can be very difficult for both writer and reader to remember at all times that when that word is used it should be taken to mean just what it was defined as and not what they had always used it to mean previously. 5 It

is as well to reiterate at this point that while over time the ‘no deal’ camp came to stand for the position of the UK’s not leaving the European Union unless there was a deal, from which perspective the rejection of a no-deal exit makes perfect sense, the original quotation in Chapter 8 comes from an earlier period in the debate where all major parties were still backing an exit and the rejection of no-deal meant a commitment to a deal, which had not yet been offered and might never be. 6 Note also, that even if this incoherence were missed, the implication that a deal must be accepted would be revealed at the next stage of analysis, and the argument would have to be reconsidered in light of that.

Fig. 11.5 A scheme for the deep linguistic analysis of arguments

188 11 CAPNA—The Comprehensive Assessment Procedure …

11.4 Informal Argument Semantics

189

The three varieties of definitional fallacy discussed in Chapters 8 and 9 were Leonard Nelson’s Philosophical fallacy, where the original definition of a word is still implicit in the new definition; G.E. Moore’s Naturalistic fallacy where a word is taken to be the same thing as that which is predicated of it; and the various types of gerrymandering of meaning which can be included under the banner of persuasive definition—an unjustified redefinition of a word to suit the current purposes of whoever is doing the redefining. These are the three listed in Table 10.1 as fallacies of definition, if there are others, they are as likely as not to appear at this stage of an argument analysis where all attempts at definition of terms are put under close scrutiny. The second level of deep language analysis is referred to simply as concept analysis, but includes a number of different errors which are the result of linguistic misunderstanding and misconception. These were also described in detail in Chapters 8 and 9. These included errors over paronymous words, where cognates are assumed to be linked by more than a shared etymological route. In argument analysis, any inference which relies on the transition from one word to an apparently related, but not identical, form of the same root, is immediately under suspicion, and the non-inflected nature of English makes it even more vulnerable to such sleights of tongue. In the oft-repeated example, there is no difference between ‘pleasure is good (adjective)’ and ‘pleasure is good (noun)’ to the listener. It is perfectly possible to read that which is ‘unreadable’; and it would be a task indeed to infer that a thing was either terrible or terrific on the basis of its causing terror. Where a similar-looking word is mistakenly assumed to have the same meaning, only in a different part of speech, a concept fallacy has occurred. There were also confusions over the use of abstract notions. Philosophically, the status of abstractions may be controversial; but for the purposes of this assessment scheme it is enough that we recognise that abstract ideas may sometimes be treated in a way too literal and too close to our understanding of notions with physical referents. This is a belief in the ability of words to describe reality and, thus, the ability of the philosopher to discern reality through the meaning of words. A misconception pointed out by Berkeley and Mauthner, as described in Chapter 9, and something I refer to as the fetishisation of language. This type of error is related to that above where the common treatment of words of the same part of speech led to incoherence. With abstracts too, there is a suspicion that their being nouns makes them susceptible to being considered things. At every step, we see how interrelated all the reasons to reject arguments actually are. When we base a claim on a definition of ‘love’ for example—and where would literature and popular music be without their constant searching for such a definition?—we are probably committing a fallacy of definition, since however we define it, our audience will judge our efforts by how well they capture the sense of the word which they already have; we may well be committing a fallacy of coherence, by assuming that ‘love’ is the kind of thing which can be described; and it seems inevitable that we are also committing a fallacy of concept, believing that because we all understand

190

11 CAPNA—The Comprehensive Assessment Procedure …

how the word is used, there must be some universally accepted definition of what it is: that ‘love’ actually refers to an abstract with some form of existence. Thirdly, there is the use of language to refer to that to which it simply cannot refer. This is the language of mysticism and, while for some it may be enlightening, comforting, beautiful, or profound, it is not susceptible of semantic evaluation, and should be rejected for that reason in an analysis such as this. For those within one tradition, there may certainly be argument both employing those terms and discussing how they should best be understood—two Taoists discussing the nature of the Tao,7 for example—where argument among mystics may continue in the language of mysticism, but any disputant from outside that tradition has the right to discard arguments so-framed as impossible of refutation or even comprehension. The two levels of deep linguistic analysis discussed so far have, then, covered the two main sources of philosophical fallacy described in Chapter 9 and pointed out by Ludwig Wittgenstein. Fallacies of definition are committed by those guilty of trying to adapt the use of language to how they would like it to be, and fallacies of concept are committed by those who have been seduced into believing that language reflects reality, or that it can be used to move beyond the human experience which has created it. For the analyst, definitions are easier to spot and to then consider, the fallacies of concept require very careful thought and may also be affected by one’s philosophical predispositions. That latter issue is not a real problem when a question is being debated though: if one discovers that the reason an argument didn’t seem quite right was because of certain conceptual assumptions made by its author, then the subsequent course of the discussion is set: possibly towards a so-called deep disagreement (Fogelin 2005) perhaps towards resolution. If, on the other hand, this careful analysis reveals that the proponent of the claim has been seduced by the vagaries of linguistic expression into giving an argument which is based upon the accidental features of that symbolic representation, or reaches outside of its communicative possibilities, then rejection is inevitable. Considered as a set of instruction for using the scheme in Fig. 11.5 these last paragraphs are perhaps annoyingly vague. The difficulty here is that in the deep semantic examination of natural language, the analyst is very much on his own. The sheer multitude and variety of possible errors or manipulations of this sort is such that no clue can be given as to how to be sure to get at them all. The infinite variety of language means that the assessment scheme must be viewed more in the light of a checklist than a precisely defined procedure. Conceptual errors due to the fetishisation of language exist: anyone assessing arguments should be aware of this and on the look-out for them. Fortunately, in very few commonplace arguments will such analysis prove necessary, and in those deeper, more philosophical, claims where it does, it will also seem natural.

7 See

A.C. Graham’s Disputers of the Tao (1989) for a full discussion of such debates.

11.4 Informal Argument Semantics

191

In Fig. 11.5, arguments rejected at both the definitional and conceptual analysis levels are redirected to possible reformation. This is held out as a possibility, but, in reality, it is difficult to see how such arguments could be rephrased so as to avoid the error they have been found to contain. That error is not so much an error of the expression of a piece of reasoning, it is more a reflection of an error of reasoning based on an error of interpretation of the medium of expression. The argument has not been badly, or deceptively, put: the argument is a bad one, the weakness of which has only become obvious when the form of its expression has been closely examined. This is something which these fallacies have in common with those of ambiguity— the entire force of the inference rests upon the meaning, the very nature, of the words which have been used in its expression. It cannot be re-expressed using other words: it is these words, and these words alone, which the argument is concerned with. Reformation is offered to the proponent of the argument, who may feel a similar point can be made with a different argument, but once a claim has faltered at this step, the analyst has no need or obligation to try to reform it and assess it once more out of generosity. The final level of linguistic analysis, and thus the final section of the entire CAPNA, is the implication analysis. This is more a question of linguistics than philosophy, but involves an equally deep consideration of the words used, beyond their surface meaning. It was mentioned earlier in the chapter when discussing the initial analysis stage that some utterances are made purely for the import of their implicatures, and the surface meaning can be safely discarded and ignored. In such cases the argument which reaches the final levels of analysis has already been reformulated to say more directly what it means. This level is rather dedicated to drawing out elements of the argumentativity of words and structures, discussed in Chapter 3, and subtler implications, suggestions, nods and winks, which have been missed at earlier levels. Where a premise or indeed the conclusion is not actually stated, only implied, the argument could not have passed through any level without its being made explicit, as there would have been no argument, but where each part of the argument is present, then each part should be examined for what it might imply or what might be inferred from it. The most obvious cases will be those involving evaluative and emotive language. It may have been a source of surprise that there is no level of analysis in the scheme monitoring non-neutral language use; there is, however, nothing objectionable about such language per se. The issue with evaluative language is the evaluations that it brings with it as implicature. This, then, is the obvious place to start. Any language which is found to be (in the eyes or ears of whoever is considering the acceptability of the argument) designed to provoke an emotional reaction or to carry with it an implicit judgement, is carefully examined. Such language has baggage, and it is only when that baggage has been unpacked and spread out neatly on the bedspread that the full argument can be assessed. It is for this reason that in the diagram of this scheme, there is no rejection of arguments because of their implicatures—implicating is not an error, it is a feature of speech. What is important is that the full argument with all

192

11 CAPNA—The Comprehensive Assessment Procedure …

relevant implicatures made explicit is presented for assessment from the beginning of the scheme, with nothing hidden or assumed. As was the case with other elements of the linguistic analysis, it is up to the assessor to look at the language used in the argument and see what implicatures there are, what argumentativity is present, and, crucially, to decide whether or not it is significant to the argument. There is a danger, one might argue, of an infinite process where the newly discovered implicatures are added to the argument and then when that fuller form is brought once more to the same stage of analysis, yet more implicatures of those implicatures are discovered and so on ad infinitum. There are two reasons why this worry should not concern us overly. The first is the point already made that the implicatures must be of significance to the argument, something which the assessor should be prepared to justify if he later uses one as the basis for its rejection. The second, as discussed with reference to the process of initial analysis, full scheme analysis, and that analysis as an initial stage of yet more analysis, is that once a ‘reasonable’ (with the agreement of all sides) attempt at finding a cause to reject the argument has been made, and has proved fruitless, there is an obligation to accept the argument, publically at least. Everyone has the right to withhold acceptance in order to think carefully, but not the right to deny it forever. Also, as Wilson and Sperber point out when discussing the retrieval of implicatures: ‘With certain types of utterance – for example a sacred text […] – a hearer might be willing to devote a lifetime’s effort to the interpretation process’ (1986: 262), so with arguments, the effort one is willing to make in order to find fault will depend on how important one deems the acceptance or rejection of the conclusion to be. Where significant implicatures are found, they are added to the full argument and the new whole is assessed from the beginning. Where none is found, there is no further obstacle to acceptance, and the argument passes into the green triangle of respectability. This chapter ends with a brief encapsulation of the procedure for applying the IAS in the assessment of expressed reasoning. I have said above that the infinite variety of natural language structures makes it impossible to provide a fully-elaborated stepby-step set of practical instructions, but the Table 11.1 gives an outline at a glance of what questions the analyst will want to address, without referring again to the likely answers to those question, nor to what is to be done with them, as this has been discussed in detail above. In all cases, the questions refer to all the language used in the most complete form of the argument, premises and conclusion, which is available.

11.4 Informal Argument Semantics

193

Table 11.1 Suggested procedural questions for Informal Argument Semantics Level of analysis

Sub-levels

Procedural questions

Vagueness

Under-determination

Is the language precise enough to be meaningful? Is the language precise enough for the purposes of the argument?

Obfuscation

Is the language unnecessarily awkward? Have words been used outside of their accepted semantic range? Can the language be understood and put more plainly? Does using the simpler version make a difference to the argument?

Ambiguity

Amphiboly

Does the argument contain structures which can be interpreted in different ways? Is the correct interpretation obvious from the context?

Equivocation

Does any term feature more than once in the argument? Is the meaning of the term exactly the same in all instances?

Ambiguous claim Coherence Deep analysis

Can the conclusion be understood in more ways than one? Is the argument formed of semantically coherent sentences?

Definition

Is there a definition in the premises? Does it attempt to redefine a word for its own purposes? Can re-definition be justified? Does the argument equivocate on the new and old meanings? Are new meanings being prescribed to fit a preferred conception?

Concept

Does the argument rest on relations between cognates? Have the cognates been confused? Does the argument take linguistic expression as a reflection of reality? Does it rely on the reality of abstractions? Does it go beyond the bounds of linguistic expression? Is the language of the argument mystical?

Implicature

Are there emotional or evaluative terms used? What implicatures do these and other terms have? (continued)

194

11 CAPNA—The Comprehensive Assessment Procedure …

Table 11.1 (continued) Level of analysis

Sub-levels

Procedural questions Are those implicatures of significance to the argument?

Chapter 12

Application of the Procedure

In this, the final chapter, I take a number of examples of apparently argumentative texts and analyse then following the procedure I have described above. These cases have been chosen in order to illustrate how the procedure can be practically applied, and are not, therefore, being presented as, in any sense, a random sample of such texts. This chapter is not designed to prove anything: not that the conceptions of argument and language upon which the CAPNA is based are correct; not that the procedure captures every conceivable argumentative eventuality; not even that it represents the best way to assess arguments and evaluate the language used in making them. The examples dealt with here illustrate merely how the procedure works in practice, and if deficiencies in its structure are made apparent, then they will at once become suggestions for its improvement. The discussion of each example will include some comments on its procedural appropriateness and the soundness of its reasoning, but the primary focus will be on the analysis of language. It is also worth repeating an important point about the initial analysis stage. At that stage of assessment, the analyst has in his head all his knowledge and experience of dealing with arguments; this may allow him to immediately recognise a fallacious pattern of reasoning and, as it were, jump to the relevant section of the scheme at once in order to justify its rejection. The procedure is not to be followed slavishly when the outcome is already obvious. Equally, an inexperienced analyst must only discover whether there actually is an argument contained in the text in order to begin to apply the steps of the three main stages of the CAPNA: Process, Reasoning, and Language. It can be assumed, then, that the examples chosen are not likely to be immediately rejected due to obvious flaws: the point of the demonstration is to show how the scheme can catch the difficult-to-see errors in the argument, and, indeed, those which may not be universally considered to be errors at all, but are points of potential rejection, depending on considerations peculiar to the audience. That might concern strength of reasoning and what level of evidence is thought sufficient in one case by one person, or it might concern the worldview implications of a particular expression, unacceptable to some, incontrovertible to others. © Springer Nature Switzerland AG 2021 M. Hinton, Evaluating the Language of Argument, Argumentation Library 37, https://doi.org/10.1007/978-3-030-61694-6_12

195

196

12 Application of the Procedure

Finally, in these introductory comments, and as a corollary to what I have just said, although the diagrammatical representation of the scheme uses the fields ‘Accept’ and ‘Reject’, the purpose of the procedure is not so much to give final judgements as to show where an objection can be raised and, thus, to lead to an extension of the argument, rather than necessarily putting an end to it. Rejection at any stage is an invitation for the proponent or a supporter of the proposed standpoint to reformulate and improve the argument. This reflects the range of modes of argumentation which I described in Chapter 5, not all of which are confrontational or even dialectical in nature. The procedure for assessment proposed in the previous chapter is equally applicable as a method of analysing one’s own arguments or those one is inclined to agree with as well as those to which one is neutral or opposed. Indeed, the point of analysis ought to be to allow one to form an opinion rather than to look for ways to cement the opinion one already has. With that in mind, the CAPNA is likely to be of most use in deciding what to do with arguments which raise some doubt or discomfort which cannot be immediately resolved, and in showing others where their arguments have gone wrong or could potentially be attacked.

12.1 Example 1—Denial of Abortion The text is taken from a report in The Guardian online newspaper, first published on 23rd February 2018 (Gentleman 2018). The article cites Ruth Halperin-Kaddari, the vice chair of the United Nations committee on the elimination of discrimination against women, discussing a report conducted by that committee into abortion rights in Northern Ireland, which, until October 2019, had a far more restrictive legal framework than the rest of the United Kingdom.1 There are two direct quotations. The first is: ‘the situation in Northern Ireland constitutes violence against women that may amount to torture or cruel, inhuman or degrading treatment’, and the second is: ‘Denial of abortion and criminalisation of abortion amounts to discrimination against women because it is a denial of a service that only women need. And it puts women in horrific situations.’ Although the original report may have carried more information, an initial analysis allows us to dismiss the first statement as just that, a statement which is not supported by any argument. The assertion that the denial of abortion rights equates to ‘violence against women’ is obviously one that could be argued over as the use of the word ‘violence’ is at least unconventional, but no argument for that standpoint is given here. The second statement immediately looks more like an argument since it features two clauses joined by the subordinator ‘because’. There is a standpoint, that ‘denial of abortion and criminalisation of abortion amounts to discrimination against women’ and one premise ‘it is a denial of a service that only women need’. The final comment 1 See https://www.amnesty.org.uk/abortion-rights-northern-ireland-timeline for an overview of developments.

12.1 Example 1—Denial of Abortion

197

‘And it puts women in horrific situations’ is difficult to analyse because it is not clear whether it is another premise to the main argument, that denial of abortion is discriminatory, or the beginning of a second argument against the existing legal arrangements. I would lean towards the latter interpretation, so it is left out of the assessment, but this typifies the difficulties faced when dealing with arguments as actually expressed. Whether or not we feel it necessary to reconstruct the argument by adding a second premise will depend upon the type of reasoning assessment we want to carry out: what we already have looks like an archetypal argument form for placing on the Periodic Table, but for a more familiar logical analysis it would require a Toulmin style warrant. That would give us: P1:

Denial and criminalisation of abortion is a denial of a service only women need. P2: Denial and criminalisation of a service only women need amounts to discrimination against women. Conclusion: Denial and criminalisation of abortion amounts to discrimination against women. The first premise has required only the elaboration of the pronoun ‘it’ from the original statement. The new second premise is relatively unproblematic since it is constructed from the same vocabulary and structures as the original speech. There are some issues which might be raised immediately, but we clearly have an argument which can be moved on to assessment in the full scheme. Process: Given that this argument is being addressed publicly to an audience which cannot immediately respond—I take the audience to be the government of the United Kingdom and, to a lesser extent, the general electorate who read the comments in the newspaper—the mode is rhetorical, and, since it represents condemnation of the status quo, antagonistic. As an interview with a journalist by a representative of an international organisation which has carried out research on the matter in hand is an appropriate forum for such a mode of argument, and as we have established that there is a standpoint and an argument given in its support, there is no reason to reject the argument at this stage. Reasoning: The argument as set out above is valid. So, are the premises true? This argument is interesting precisely because this is a difficult question to answer. At first sight, the first premise looks convincing, since only women can have abortions, so, presumably, only women can need them. The second premise is clearly a question of definition, suggesting at once that linguistic assessment will be important. If we take the Wagemans approach, we can identify the original statement as of the form: ‘a is X, because a is Y’. This makes it a first-order predicate argument. The further categorisation does require a closer look at the language, but if we assume that claiming that the law is ‘discrimination’ is a value judgement, we get the code 1 pre VF, which is named in the table as ‘Cr’ argument from criterion. This in itself is a sign that linguistic analysis may be needed, since interpreting whether or not a criterion has been met is likely to involve careful consideration of certain definitions. It is also worth noting here that the Cr argument can be said to imply an Ev one— from evaluation—which has the form 1 pre PV, and would argue that the law should be changed because it is discriminatory.

198

12 Application of the Procedure

Language: Following the list of questions presented above in Table 11.1, we begin with vagueness: under-determination and obfuscation. Is the language precise enough to be meaningful? Is the language precise enough for the purposes of the argument? The denial and criminalisation referred to are those found at the time in the province of Northern Ireland, so that is clear. There is, however, a potential problem with the phrase ‘a service only women need’. What does it mean to say that a woman needs an abortion? The only situation where abortion was legally available in Northern Ireland was when the mother’s life or health were seriously threatened, so ‘need’ here cannot refer to such situations, unless the argument is connected only with practice and not with the law. That is not the understanding of either the author of the article or Grainne Teggart of Amnesty International, who is quoted in the same piece as saying ‘The UN committee is very clear that it is the UK government which is responsible for ensuring that our laws are in line with the state’s international human rights obligations’. If ‘need’ is understood to mean require in order to live fully, comfortably, as one wishes, it becomes less clear whether ‘only women’ require pregnancies to be terminated, since fathers are also legally responsible for their children and have their lives changed by their appearance, albeit generally to a much lesser extent. If that is the interpretation used, then the argument collapses. It may be that the speaker meant ‘want’ but shied away from using the word as it might sound as though she were supporting all abortion on demand; it would have seriously damaged her argument as her standpoint would have been reduced to ‘not giving women services they want is discrimination’. It would also make the ‘only women’ element untrue. The most generous interpretation of ‘need’ here which allows the argument to move forward is that she means it is a ‘service’ which can only be performed on women, that only women can have done to them, that only applies somehow to women. That is beginning to look rather weak though. One could see this as ambiguity in the word ‘need’, but it is not so much the case that there are multiple meanings from which we must choose, more that it is hard to see how exactly ‘need’ could be defined at all in this context. The difference is a subtle one, and overlap between the categories is both to be expected and to be accepted as normal. Is the language unnecessarily awkward? Have words been used outside of their accepted semantic range? The use of the term ‘service’ is questionable here. Does it refer only to the termination procedure? If not, what else is it supposed to cover? Is it being used euphemistically in order to avoid connotations with the unpleasantness of the actual operation? This is an example of language use which is worthy of note, and might suggest a clearer rewording is necessary, but it is not fatal to the argument as a whole. Next ambiguity: amphiboly, equivocation, ambiguous claim. Does the argument contain structures which can be interpreted in different ways? No. Does any term feature more than once in the argument? Is the meaning of the term exactly the same in all instances?

12.1 Example 1—Denial of Abortion

199

Yes, and yes. Need has been discussed as being vague rather than having competing meanings, although the line between the two is not easily drawn. The only other content word repeated in the original quotation was ‘denial’, and there is no reason to see a slide in its meaning. Other repetitions come about as a result of the warrant premise, where the meanings are copied, as it were, from the original statement. Can the conclusion be understood in more ways than one? No, within the context, the claim is unambiguous. The issue with need is confined to the premises. Whether the claim tacitly implies further claims about policy matters is a different issue. Coherence: Is the argument formed of semantically coherent sentences? No. Here we meet the next clear problem. The premise states that: ‘Denial and criminalisation of abortion is a denial …’. Clearly, denial of abortion is a denial, and the addition of criminalisation only confuses the matter. One might argue that that which is criminalised is also denied, but, in fact, this phrase is bringing in a second argument, since the denial of the service and the criminalisation of the procedure are not exactly the same thing. This means a rephrase is necessary. One way to do it would be to simply remove the ‘and criminalisation’ part from both premises and conclusion, in spite of the fact that it is only in the first premise that it is semantically problematic. This is attractive since it does not require any additions which may not be supported by the proponent of the argument. However, such a change would also disturb the logic of the argument. That would be better preserved by adding ‘and criminalisation’ to the predicate of premise 1, giving us: P1: Denial and criminalisation of abortion is a denial and criminalisation of a service only women need. Or, with the changes suggested to avoid vagueness added: P1: Denial and criminalisation of abortion is a denial and criminalisation of a procedure only women can have. This is obviously a little awkward in terms of natural speech, but since the speaker brought in both denial and criminalisation, both need to feature at all stages. The argument is already looking shaky, but has not been rejected. Finally, the deep linguistic analysis, beginning with Definition. Is there a definition in the premises? Does it attempt to redefine a word for its own purposes? Claiming that denial of a service needed by women is discrimination against them does seem to be at least testing the definition of discrimination. The phrase ‘discrimination against’ is generally used to mean different, worse treatment for one group as opposed to another. It isn’t clear that there can be discrimination against women where only women are involved. Can re-definition be justified? Possibly. We might feel that this expansion of the meaning of discrimination is justified since it refers to a wider societal unfairness rather than differential treatment in one area. Are new meanings being prescribed to fit a preferred conception?

200

12 Application of the Procedure

The suspicion here is that, yes, they are. The arguer wants to be able to claim that the law is discriminatory: not just wrong or unjust, but discriminatory. An opponent would be able to claim that this use of the word ‘discrimination’ is a subtle attempt at persuasive definition and should be rejected. Does the argument equivocate on the new and old meanings? One of the interesting points about equivocation, which was noted in the discussion around the word ‘poverty’ in Chapter 8, is that the person making the argument may be able to claim that there is no equivocation and they have used a word consistently in cases where an audience may be expected to understand that word in a different sense. Here, while the reasoning does not rest on using the same word with two different meanings, it can be expected that a public audience would understand ‘discrimination’ according to its more standard meaning, while the argument is only valid if the newly expanded, slightly unusual sense made explicit in the unstated second premise is actually the meaning being employed. As with the poverty example, the arguer wants to use this particular word because of its connotations, but those connotations belong to the original, not the adapted meaning. To recap: there was a little semantic untidiness, which has been cleared up; there is uncertainty over the meaning of ‘need’, it must refer to something physical, in order to exclude men, but cannot refer to serious medical needs, as in this case the procedure was not denied. There is also doubt over the application of ‘discriminate against’ to situations which involve only the group being discriminated against. These two issues are certainly enough to prevent acceptance of the argument. This strikes me as having got to the heart of what it was which seemed unsatisfactory about the original argument, and could be the end of the analysis, pending clarification of how need is to be understood, and perhaps a rewording without the use of ‘discrimination’. There does not seem to be any call here for a conceptual analysis, but the final stage of Implicature is worth pursuing as it reveals more about the interconnectedness of the earlier stages. Are there emotional or evaluative terms used? What implicatures do these and other terms have? Are those implicatures of significance to the argument? Defining what exactly are emotional and evaluative terms is a practical impossibility, but it is not necessary: we are interested generally in what terms imply, rather than whether or not that implicature should be called emotional or evaluative. The key term here is ‘discriminate against’; it is presented as a factual judgement but is, in fact, a value judgement, the implication is that the law denying abortion to women in Northern Ireland is wrong, is unfair, and should be changed. The word ‘need’ implies that what is being denied is a necessity, something women cannot do without. That increases the importance of the problem, but also leads to a questioning of the way in which they cannot do without it. These implicatures were revealed in the search for fallacies of vagueness and definition. The definition questions also partially revealed another important inference which can be made from the attempted adjustment of the meaning of ‘discrimination’. If it is true that: P2:

Denial and criminalisation of a service only women need amounts to discrimination against women then, presumably, it is true that:

12.1 Example 1—Denial of Abortion

201

P2 : Denial and criminalisation of a service only one group in society needs amounts to discrimination against that group. The importance of making this explicit is that it encourages the audience to look for examples which might apply to other groups and decide whether the principle really holds for all groups. If it does, then the expanded meaning of ‘discrimination’ can be put to use and the principle can be applied to other injustices, not only the specific situation over abortion rights in Northern Ireland. If it does not, then an explanation will be needed as to why women are to be treated as an exception.

12.2 Example 2—No Such Thing as Society The second example centres around a well-known and controversial statement and a response to it. Thus, the assessment will be a little more complicated, as it will take in both the original argument and an attempted refutation. In 1987, the then Prime Minister of the United Kingdom, Margaret Thatcher, gave an interview to Douglas Keay, for publication in a popular British magazine, Woman’s Own, in which she made the famous statement that: ‘There is no such thing as society’ (Thatcher 1987). This was neither a slip nor a casual soundbite: she repeats the point and it comes as part of a long and careful explanation of her views on a number of ‘social’ issues. It is worth noting that, whatever opinion one may have of her politics and her legacy in the UK and elsewhere, Margaret Thatcher was both an exceptional speaker and a politician with a reasonably consistent philosophy with which she could back up particular policies: something which cannot be said for any of those who have succeeded her as Prime Minister, from either the Conservative or Labour parties. By calling her an exceptional speaker, I do not mean that she was a great orator, although she was quite capable, but rather that she was able to speak in interviews using complex and coherent sentences to actually make sophisticated points, rather than attempting to survive questioning without making an error, which is often the approach taken by political interviewees. As such, Margaret Thatcher’s archive of pronouncements is a gift to argumentation researchers who may struggle to find any arguments in the words of today’s world leaders. The passage from the interview which attracts the most attention runs thus: I think we have gone through a period when too many children and people have been given to understand “I have a problem, it is the Government’s job to cope with it!” or “I have a problem, I will go and get a grant to cope with it!” “I am homeless, the Government must house me!” and so they are casting their problems on society and who is society? There is no such thing! There are individual men and women and there are families and no government can do anything except through people and people look to themselves first.

This is followed a few lines later by: There is no such thing as society. There is living tapestry of men and women and people and the beauty of that tapestry and the quality of our lives will depend upon how much each of us is prepared to take responsibility for ourselves and each of us prepared to turn round and help by our own efforts those who are unfortunate.

202

12 Application of the Procedure

There is clearly at least one argument being made here, but some reconstruction is necessary, as there is actually a chain of inferences implied. The ultimate conclusion appears to be the final part of the second quoted section, that quality of life depends on each of us taking responsibility for ourselves and making an effort to help others. This is true because only individuals can act, ‘no government can do anything except through people’. The argument can be reconstructed in the following ways: P1: There is no society, only individuals. P2: Only that which exists can act to solve people’s problems. Conclusion: Only individuals can act to solve people’s problems. Or, as a sentence without the unnecessary warrant: Only individuals can act to solve people’s problems because only individuals exist. In the Wagemans notation this would again be of the form ‘a is X because a is Y’, and since both statements would be factual, the argument would be one of correlation: being able to act correlates with being. The second part of the argument runs: P3:

To ensure quality of life, someone should act to solve problems and help people. P4: Only individuals can act to solve people’s problems. Conclusion: To ensure quality of life, individuals should act to solve people’s problems. For the time being, let’s take these two steps into the main assessment procedure. Process: The context of this argument is a little different from a political speech. The magazine Woman’s Own in the 1980s was a conservative publication aimed at mothers and housewives rather than a hotbed of serious discussion on current affairs. The questions put to the Prime Minister were not in any sense aggressive or probing, but simply an invitation to talk. Because of this, the mode of argumentation is more cooperative and explanatory than rhetorical or confrontational. Mrs. Thatcher is being invited to explain her policies and she does so by setting out how they are grounded in her overall view of society. With that in mind, her responses are appropriate. Reasoning: One of the problems with reconstructed arguments is that they tend to be logical. If the author of a text does not explicitly make a mistake with the construction of the argument, then the analyst can hardly add one, and in the spirit of charity, it is usually possible to make some kind of valid argument from the fragments of text, even if it must rely on very questionable premises. In this case, the train of thought in the original utterances is quite logical so a valid reconstruction was not difficult. It should be noted that both the steps rely on deductive inferences: the conclusions follow necessarily from the premises. The question, then, once more, is over the truth of the premises. Taking the second stage first, P4 is simply the earlier conclusion; P3 is, I think, a relatively uncontroversial moral principle. The exact wording reflects the way it is put in the original, but the basic idea that to improve life for all, we should try to solve our problems and

12.2 Example 2—No Such Thing as Society

203

help one another when necessary, is likely to be accepted by most audiences. P2 is a warrant premise which is close to being a tautology—that which is not, cannot—and might inspire doubt in only the most refractory philosopher. So all comes down to whether or not P1 is acceptable. In a sense, it is clearly false to say that society does not exist, since it is all around us, and we are all part of it. This might lead us to reject the argument as unsound, based on a false premise; and that is what opponents of Mrs. Thatcher have generally done. The following comes from a short essay2 discussing the famous line: Perhaps the first important thing to note is that these are semantic games. Society is a “collection of individuals,” even formally defined. […] the moment two or more individuals establish any sort of cooperative relationship with each other, the inevitable result, by definition, is a society. To say that there is no such thing as society is demonstrably false. Humans are born in groups, raised in groups, work in groups, play in groups, defend their interests in groups, and die in groups. (Huppi, n.d.)

There are two things to note about this attempt at refutation. Firstly the writer talks of ‘semantic games’, a phenomenon discussed in Chapter 6, in an apparent dismissal of the importance of the meaning of words. Secondly, the writer makes something of a fool of himself by suggesting that Mrs. Thatcher was not aware of the fact that human beings live in groups. Since it is obvious, from the rest of the interview, if nowhere else, that she is aware of this fact, it is equally obvious that she did not mean to say that groups of people do not exist. The problem with Huppi’s argument, then, is that it is a variety of Straw Man. It is a more forgivable instance than most, however, because to understand what the original argument actually was requires a certain generosity and seriousness of linguistic analysis; still, when a serious person bases an entire political philosophy on a claim which at first sight appears to be obviously false, it makes sense to look a little more deeply into what was actually meant. This level of generosity and seriousness is frequently lacking in political debate, which is why it is so full of Straw Man arguments. Before completing that serious linguistic analysis, it is interesting that Huppi also accuses Thatcher of having committed a fallacy of composition. The belief that humans are autonomous individuals leads to a logical error called the “fallacy of composition.” This fallacy holds that aggregate behavior is the same as its individual parts. This is hardly true. Atoms may have certain distinct properties, but they rarely appear alone in nature; they almost always form molecules. […] The same holds for individuals in society. Most of us would agree that individuals are highly self-interested. But this inner trait manifests itself in different outward behaviors, depending on whether the individual is alone or in a group. […] There are numerous other examples of social behavior that cannot be explained on an individual level.

Fallacies of composition are committed when a claim is made that something which is true of a part of something is also true of the whole. That would mean Thatcher 2 The

essay appears to have been written by a Tom Huppi and made available via his website at http://www.huppi.com/kangaroo/L-nosociety.htm.

204

12 Application of the Procedure

would have to be making a claim that something was true of society because it is true of the individuals who make up society. It is obvious that she makes no such claim, since she claims that one exists and the other doesn’t. What Huppi seems to want to say is that individuals on their own and individuals in society behave differently, that people naturally come together in groups which have different properties from those people as individuals. This is a reasonable point, but it does not address anything Thatcher said. Language: This example has been chosen largely to demonstrate the flexibility of the approach I have outlined, and to illustrate that the theoretical framework for understanding the relationship between language and argument is not dependent upon the ability of certain questions to fully analyse every example. In this case the linguistic analysis is an attempt to save an argument which, at first glance, looks easy to dismiss. Since one premise is obviously false on a simple interpretation and the source of the argument is not a naïve or frivolous one, the resources of the linguistic analysis scheme can be employed in trying to reach a more accurate interpretation of the origin text. The starting point of the analysis, then, is that the simplistic understanding, the one shown in the critical essay described above, is in error, and the premise, P1, should be understood differently. By asking why the premise is frequently misunderstood, we can discover, perhaps, the intended meaning. Perhaps the problem is vagueness. Is the language precise enough to be meaningful? Is the language precise enough for the purposes of the argument? Clearly, when a term is not understood as intended, the most likely culprit is a lack of precision. In saying that there is no society, when obviously we live in social groups, what is meant by ‘society’ here? Have words been used outside of their accepted semantic range? Can the language be understood and put more plainly? Does using the simpler version make a difference to the argument? The term ‘society’ has been used a little unconventionally, but without confirming what it is supposed to mean here, we cannot say whether it is being semantically misused. To put the point more plainly, again, we must look carefully at what the term is being used to mean. This, we shall return to in a moment. Ambiguity is not a feature of the original argument, although it may well be a feature of the counter-arguments, which say that society does exist because people live in groups, when this was clearly not the intended meaning in the original. The argument is coherent—societies are capable of existing or not existing, individuals are capable of acting. Is there a definition in the premises? Does it attempt to redefine a word for its own purposes? Can re-definition be justified? There is a sense in which a re-definition is taking place. Rather than an attempt to shift the meaning of the word, however, this is an attempt at precisification. The suggestion is that ‘for the purposes of this argument, in the context of what we are discussing, there is no such thing as society’. In order to decide if this move is justified, we need to examine the justification for it. P5: P6:

In this discussion, we are concerned with the actions of agents. Society is an entity with no direct agency.

12.2 Example 2—No Such Thing as Society

205

Conclusion: For the purposes of this discussion, there is no such thing as society. Thatcher is, in fact, accusing her political opponents, and much of the population in general, of committing at least one of two fallacies in their own appeals for ‘society to do something’. Firstly, and ironically given the above attack on her position, we could characterise it as an accusation of a fallacy of composition. Since individuals, as the elements of society, have the power to act, so socialists believe that society, as a whole, has the power to act. Fallacies of composition are not always clear cut: if something is true of enough of the individual parts of a larger whole, then we do often apply the predicate to the larger entity, but it isn’t obvious under what conditions. For instance, if one part of my car is broken beyond repair, my car is not a write-off, but there comes a point when so many parts are broken that I can say, ‘my car is broken beyond repair’. In the current case, most individuals in society, certainly healthy adults, do have agency, the question is more one of whether or not society is the kind of collection of individuals to which agency can be attributed. This is the second fallacy of which Thatcher suspects her critics: when they argue that ‘these are social problems, so society should act’ they make a conceptual error. Individuals have concrete existence, and society is an abstraction from the existence of those individuals. So Thatcher here would ask: Does the argument take linguistic expression as reflection of reality? Does it rely on the reality of abstractions? To which she would answer, yes on both counts. Action can only be taken by concrete entities: expecting society to act is to expect an abstraction to have the existent reality necessary for agency. Whether or not she is right is a matter for further argument, since certain groups, such as clubs, boards, tribunals are permitted to act in that form as that institution, and the individuals who act on their behalf are not, at least not always, held personally responsible for those acts. Society is a comparatively nebulous grouping, however, and I am inclined to think that Thatcher is correct to say that society qua society is incapable of agency. If we do accept her view, we can also say that any argument of the form ‘society should do something’ is also committing a fallacy of semantic incoherence, since societies are not the kinds of thing to which the predicate ‘should do’ can felicitously be applied. We can now return to the problem of insufficient clarity and rephrase the argument with greater precision. P1 : P2 :

In the context of agency there is no society, only individuals. Only that which exists, in term of having agency, can act to solve people’s problems. Conclusion: Only individuals can act to solve people’s problems. There are two further comments to make here. Firstly, I don’t believe that this type of argument is particularly innovative or unusual. It is common to hide the individual within the collective, and also common to expose this. Imagine a political leader saying ‘10% of the army will die in the attack, that’s acceptable’, to which his general replies: ‘10% of the soldiers will die: there is no army.’ There obviously is an army in one sense, but the context of dying in battle is relevant only to soldiers, the individuals who make up the whole. This is an analogous argument to Thatcher’s.

206

12 Application of the Procedure

Secondly, despite the complex analysis required to reach this point, a disinterested audience can spot this ‘in the context of’ definitional shift quite easily. When the well-known polling group Ipsos MORI presented 500 British adults with the bare statement ‘There’s no such thing as society’, as many as 74% disagreed, compared to 16% who agreed. When they tested reactions to an excerpt slightly longer than that cited above, the figures were rather different: 66% agreed with the passage, and 25% disagreed (Ipsos Mori 2013). In context, then, there is evidence that, at the very least, Thatcher’s denial of the existence of society as an agent was accepted as meaningful. This analysis would suggest that the controversial statement and the argument based upon it have been unfairly criticised and that, in fact, Mrs. Thatcher was arguing soundly. There is, however, a twist in the tale. Her ultimate aim, as mentioned above, was to justify her government’s policies which were supposedly designed to allow individuals to create wealth with which they would then be able to act to help themselves and others. In claiming that society cannot help, only individuals, she commits a fallacy I shall call a synonymic equivocation. The first part of the previously cited passage went like this: I think we have gone through a period when too many children and people have been given to understand “I have a problem, it is the Government’s job to cope with it!” or “I have a problem, I will go and get a grant to cope with it!” “I am homeless, the Government must house me!” and so they are casting their problems on society and who is society?

Twice in these lines, she suggests that unfortunate individuals are asking for government assistance, and once more for a grant, also, presumably, from the state. This she interprets as ‘casting their problems on society’. By synonymic equivocation I mean rather than taking two instances of the same word to mean the same when, in fact, they do not, instead taking two apparently synonymous words to mean the same, when in fact, they do not. The switch from ‘the Government’ to ‘society’ implies that they are synonymous, so, fully-expanded, her argument looks like this: P7: People expect the Government to solve their problems. P8: The Government means society. Conclusion: People expect society to solve their problems. This looks suspect at first glance. It becomes more so when we remember the way in which she employs the word society—that it does not exist as an acting agent. That is clearly not true of the Government. The Government has legal agency, and Mrs. Thatcher, as its head, did have the means and the power to act to address the problems of which members of her electorate complained. That she felt it better for individuals to so act than for governments to do so is a matter of political doctrine, but the quite deliberate shift from ‘the Government’ to ‘society’ means she was attacking a straw man version of the argument she herself had just created: these people were not asking for help from society, by her own account, they were asking for help from government; and, presumably, she did not believe that there is no such thing as that.

12.3 Example 3—The University of Life

207

12.3 Example 3—The University of Life The third example is another from the Guardian newspaper (31st May 2019), but this time from an opinion piece by Simon Jenkins, a regular contributor. The headline for the column was ‘What are our universities for? Taxpayers have a right to know’ and the article contained a rather incoherent set of comments, complaints and personal bugbears masquerading as a reaction to a report into student funding by Philip Augar. This is a text which illustrates the difficulty in extracting arguments from apparently argumentative content. A short excerpt will show the problem: Spending money on higher education is like spending it on defence. You suspect half of it is wasted, but you cannot tell which half. No causal relationship has been proved between graduates – even science ones – and national wealth. A good university is still the nearest a secular society gets to a sacred institution. (Jenkins 2019)

One paragraph containing three statements, one an analogy, with no connection between them and no arguments in support. The conclusion of the piece, however, does at least imply an argument and, as it happens to be an example of a fallacy which I have frequently discussed in the past without solid evidence that it exists, I shall examine it here. The final sentence of the article, preceded by one which bears no relationship to it, is this: ‘For most people, graduate and non-graduate, the best education takes place in the university of life, and that costs nothing.’ Since the article does not discuss the benefits of ‘the university of life’ I take this to be a premise rather than a conclusion. The article refers several times to the cost of university education and includes the line: ‘No one can say what the nation has gained for all this money’, which along with the suggestion quoted above that half of the money spent on education is wasted, allows me to reconstruct the following two-step argument: P1: P2:

Education at university is very expensive. For most people, graduate and non-graduate, the best education takes place in the university of life, and that costs nothing. Conclusion: A cheaper and better alternative to education at university is available. P3: A cheaper and better alternative to education at university is available. P4: Spending on more expensive and worse options is a waste of money. Conclusion: Spending on university education is a waste of money. For the record, I do not for one moment believe that the author of the piece would endorse this conclusion. I suspect that the ‘university of life’ statement was a throwaway line which had received about as much consideration as the structure of his article. Nonetheless, the argument is interesting in its own right. The initial stage of analysis has proved difficult because the arguer, despite being a professional opinion writer, has failed to properly express any position, much less an argument for it; but we do have an argument to take on to the next level. Process: One might object to the entire contribution from Jenkins on the grounds that it is labelled as a kind of reply to a report, but does not actually address the points in that report in any kind of dialectical fashion; or, indeed, on the grounds discussed above that it does not offer a coherent argument move. I shall interpret it as another

208

12 Application of the Procedure

example of antagonistic rhetoric, however, making a legitimate comment on public policy. Reasoning: Evaluation of the soundness of the reasoning centres mainly around our acceptance of premise 2. Premise 1 is relatively uncontroversial, universities do take large sums of money from government budgets, and studying, in many countries, puts significant financial strain on students. Let us suppose, for now at least, that we agree that we have learnt more from life than we ever did in lectures and accept premise 2. The conclusion is acceptable on this basis, so P3 is also accepted and P4 is no more than a definition of ‘waste’. Language: The flaw in this argument is more obvious than in example one. It is more difficult to apply all the questions concerning language fairly since the argument is so heavily reconstructed, but it is unnecessary since the most important question stands out clearly. Does any term feature more than once in the argument? Is the meaning of the term exactly the same in all instances? Yes, and certainly not. The key term in the argument is ‘education’, the article has been filed under that heading on the Guardian’s website. Even if P1 is a reconstruction, the article concerns, specifically, university education, and the word features in that context many times. In P2, ‘education’ refers to something gained outside of university. Is this the same thing? If we rephrase with synonyms we get: P1 : P2 :

Academic education is very expensive. For most people, graduate and non-graduate, the best learning experiences take place in the university of life, and that costs nothing.

From which nothing at all can be concluded as there are no longer any common terms. This, then, is a classic example of an equivocation. A taught, academic university course is not an ‘education’ in the same sense that experience of life is an ‘education’ though the word can be legitimately used for both. One extra point of interest is the way in which figurative language plays a role in encouraging this mistake. A second repeated term is ‘university’. The metaphorical expression ‘university of life’ suggests that there is a similarity between life experience and academic learning, when really they are very different things. Those who use the phrase tend to forget that students are living as well as studying: academic study and life experience are not two exclusive alternatives. The argument taken from the article can then be conclusively rejected since the first stage of it is founded on a clear equivocation. I am certain that the author, with a little thought would accept that, but since the ambiguity of the word ‘education’ is well established and frequently appealed to in the phrases ‘university of life’ and ‘school of hard knocks’ it is worth exposing here.

12.4 Example 4—Clear and Obvious Error

209

12.4 Example 4—Clear and Obvious Error One of the conclusions of the first section of this work was that all language can be seen as containing elements of argumentativity, even when it is not employed in an explicitly argumentative structure. A type of utterance which is intended to play a role in arguments—indeed can have no other use—but is not originally expressed within the context of an argument is the rule. Whether these are legal statutes or moral precepts, their purpose is to act as warrants allowing the conclusion of the propriety of some form of sanction from the data of some kind of act. They may also be employed as premises in arguments advising against or in favour of certain actions. The writing of rules is a particularly difficult and important type of linguistic expression because they have consequences and because they are usually inflexible: subtleties of meaning cannot be debated with an umpire, though they may be with a judge, and subtleties of meaning can result in winning, losing, or being incarcerated for the rest of one’s life. The fourth example for analysis comes from the world of sport and involves an area which has brought the necessity of drawing up new rules in a number of games— video technology. As the example to be looked at is a statement which has not been employed in a specific argument, but lies implicitly behind many conclusions, amongst them some angrily debated, the analysis skips the levels of reasoning and process and focuses only on the linguistic stage. The introduction of video technology to assist referees in association football has aroused a good deal of controversy and has taken far longer than in other sports, such as cricket, rugby and tennis. The need has been clear for some time with fans watching on television able to spot outrageous examples of mis-officiating instantly, which could not be later corrected: these included famous incidents of mistaken identity such as the sending-off of Kieran Gibbs after a deliberate handball by his team-mate Alex Oxlade-Chamberlain in an English Premier League match (BBC 2014), and examples of blatant cheating which had significant impacts, such as Thierry Henry’s goal for France which eliminated Ireland from the World Cup qualification process, despite being scored with his hand (Ogden 2009). At the same time, football is a game where referees are constantly called upon to make interpretative judgments about foul play and are expected to keep the play flowing with quick, unobtrusive decision making. Reflecting this, the international law-making body for the game, the International Football Association Board (IFAB) stresses on its website that: The aim of Video Assistant Referees (VARs) is not to achieve 100% accuracy for all decisions as there is no desire to destroy the essential flow and emotions of football which result from the game’s almost non-stop action and the general absence of lengthy stoppages. (IFAB 2019a).

This is clear enough, the question is how to achieve the right balance. In their VAR protocol (IFAB 2019b: 139), the IFAB writes as point number one that: 1. A video assistant referee (VAR) is a match official, with independent access to match footage, who may assist the referee only in the event of a ‘clear and obvious error’ or ‘serious missed incident’.

210

12 Application of the Procedure

Point two says that the referee must always make a decision first, which is then checked, and not wait for the VAR to decide for him. This is then followed by point 3: 3. The original decision given by the referee will not be changed unless the video review clearly shows that the decision was a ‘clear and obvious error’. The fact that the phrase ‘clear and obvious error’ is both repeated and placed in inverted commas shows that it is to be considered crucial to the operation of the rule. In reality, point 3 has often been ignored by match officials, and even questionable errors (errors in the opinion of the VAR, but contested by commentators, pundits and participants) shown on the video recording have been corrected. This has led to widespread dissatisfaction amongst fans and professionals alike as the often marginal infringements which have been spotted by VARs have been the result of debateable interpretations of the laws of the game. An argumentative analysis of the language of the clauses in the protocol can illustrate why this was, in fact, inevitable. Beginning with the vagueness questions for clause one: Is the language precise enough to be meaningful? Is the language precise enough for the purposes of the argument? To the first question the answer would be yes, but to the second, probably no. What exactly is a ‘serious’ incident? Well, the clause contains four situations in which the VAR can be involved, although the phrasing suggests that that is not intended as an exhaustive list of serious incidents, and no more is said in explanation. What exactly does ‘clear and obvious’ mean in practice? If the situation was so clear, why was the error made in the first place? The description needs to be clarified by stating to whom the error is clear and obvious, since it was neither to the referee nor his on-field assistants. One is left to assume, given the background to the introduction of the technology, that the errors in questions were those which players and spectators were aware of, but, for some reason, officials had made and not realised. This clause from the protocol should have been rejected as unworkable at this stage, purely on the basis of a fallacy of vagueness. The situation is worsened, however, when we look at it in conjunction with clause three. If we ask the same questions of clause three, the answers are more positive. There is no mention of serious incidents, and here the ‘clear and obvious’ is less vague since ‘video review clearly shows’ can be understood as meaning the error is ‘clear and obvious to the VAR’, or indeed to anyone who has seen the review. Any error which is clear in the review can be corrected by the video referee, and, crucially, since it is also clear to the audience at home and all the players and officials after the match, it must be corrected. The results of the ambiguity analysis when the clauses are taken together will be obvious: Does any term feature more than once in the argument? Is the meaning of the term exactly the same in all instances? Yes, and no. In the implied meaning of the first clause, a clear an obvious error in applying the off-side law in football would be to allow play to continue when a player is five yards in front of the last defender, but not if he is only five inches in front: that would be a marginal call. However, in the meaning of clause three, a

12.4 Example 4—Clear and Obvious Error

211

slow-motion replay can reveal very clearly and obviously that a player who is only five inches off-side is, in fact, off-side, and should be penalised. In short, what is clear and obvious in the video replay is not what was clear and obvious without the video replay. Taken as an argument to determine what the officials at the match should do, the protocol commits a fallacy of equivocation. This is the reason for the dissatisfaction of fans, although they can see (in most cases at least) that the decision of the VAR is the correct one, by the letter of the law, and they can see it clearly, they do not consider the error that was made by the referee to have been a clear and obvious one, because in fact it was extremely marginal and could not have been detected with the naked eye. The ‘in most cases’ of the last paragraph is also important. The words ‘clear and obvious’ are also ambiguous in the sense that what is clear may be what can be seen, or what can be understood from it. One of the reasons that officials do make ‘clear and obvious’ mistakes in the meaning of clause one is that they did not see the incident, a problem which does not apply to the VAR; that does not mean though that any incident which can be seen clearly is equally clear in its interpretation. The result is that even when the VAR has a clear view of an incident from a number of angles, it may not be clear what the correct decision is: such is the nature of the game, a contact sport containing legitimate aggression. The rule would suggest that in cases where interpretation is required, the on-field referee’s interpretation should be the one maintained, however, if the VAR believes that he can see details which his colleague could not, then it becomes a matter of judgement whether the difference in their opinions is a difference of interpretation or a difference of evidence, clearly seen in the replay. This is all compounded by the phrase ‘clearly shows’ in clause three: the video is supposed to clearly show a clear error, but it isn’t clear if clear means the same thing as clearly in this sentence; and how it might show a clear error, but not clearly show it, or clearly show an error which is not clear, are mysteries untouched. The conclusion to this might be that IFAB ought to consult with a linguist before they introduce any more new changes to the laws of the game: although from their recently introduced change to the handball law which makes it an offence if a player ‘touches the ball with their hand/arm when: the hand/arm has made their body unnaturally bigger’ (IFAB 2019b: 104), it seems they are still unaware of the need. This case provides an example of semantic shifting which is not due to an attempt at manipulating words or facts to prove a point, but is caused by a change in circumstances: part of the argument appeals to the perspective of the audience in one situation, and part of it to their perspective in a very different one, and it is not only the degree of evidence available which has changed, it is the meaning of the language used to describe that evidence.

212

12 Application of the Procedure

12.5 Example 5—A World Without Grammar The final example for analysis in this chapter is, fittingly, a text which attempts to make an argument about language and its relationship to critical thinking. I say attempts to, because Douglas Rushkoff’s (2013) article is one of those which appears to want to argue for something, but never really states what exactly. It is also only fair to point out that practically everything Rushkoff says in his thankfully short piece is completely wrong, from his apparent belief that grammar is exclusively a part of written language, to his assertion that for most jobs good writing skills are essential. Although the article contains no expression of what might be called a standpoint, we can assume from the number of reasons given for the importance of grammar, that the author wants to persuade us that grammar, the use, teaching and learning of it, is important: the pedagogical aspect of this assumption is justified as the piece was part of a debate entitled ‘Is our children learning enough grammar to get hired?’ Since no standpoint is put forward and no evidence is given in support of the various claims made, it would take the most generous of analysts to even move this text past the preliminary stage and into Process analysis, where it would undoubtedly fail anyway. It is discussed here, therefore, not so much as an example of how the assessment procedure can be used in practice, but rather more as an example of an argument concerning some of the topics of this book and confirming some of the conclusions I shall reach below. Rushkoff begins by suggesting that grammar is a facet of the printed word and is not to be found in ‘YouTube videos’, and then proceeds to explain why it is still important. Presumably he accepts that YouTube videos do contain language, at least sometimes, rather than unconnected signs, so we may suspect that by ‘grammar’ he does not actually mean ‘grammar’ at all, but something else, which we shall perhaps uncover as we read. The reasons why we should not ‘cast off’ grammar are many (I have numbered them) and run as follows: without grammar, we lose the agreed-upon standards about what means what [1]. We lose the ability to communicate when respondents are not actually in the same room speaking to one another [2]. Without grammar, we lose the precision required to be effective and purposeful in writing [3]. […] Only the writer skilled in grammar is entrusted with representing a company in a letter or an e-mail [4]. Only the entrepreneur who can persuasively express a new idea in writing can craft a business plan that will win the faith of partners and investors [5]. […] Even a poorly constructed tweet reflects a poorly constructed thought [6], while grammatically lacking e-mail messages have become the hallmark of password phishing scams [7]. Without command of grammar, one can’t even truly read, much less write [8]. […] an employee who can write properly is far more valuable and promotable than one whose ambiguous text is likely to create confusion, legal liability and embarrassment [9]. Moreover, a thinking citizen deserves the basic skills required to make sense through language, and to parse the sense and nonsense of others [10].

Numbers 4 and 9 are more or less the same thing, but that leaves as many as nine distinct reasons to consider grammar important. Putting this multitude systematically through the entire assessment scheme would be a long and tiresome process. Here,

12.5 Example 5—A World Without Grammar

213

then, we can employ the elements of the procedure as expert analysts, immediately turning to the relevant stages to make pertinent comments on each claim. [1] This statement looks to be correct: grammar is obviously part of meaning. As a premise to an argument in favour of teaching grammar, however, it commits an equivocation fallacy. The grammar essential to getting meaning from language is not the grammar taught at school—if it were, there would be no linguistic communication without formal education, which is clearly not the case. [2] If ‘grammar’ is understood here as linguists understand it, then this argument is one of the curious, though common enough, type which reach a finely detailed conclusion when a more general one would do. Without grammar there would be no communication with language, whether we were in the same room or not. The author is suggesting that grammar is essential for writing, and by implication, not for speaking. Suppose then that he means ‘schoolroom grammar’ rather than linguistic structure. This argument falls down just like the first: while literacy does need to be taught, it isn’t necessary to teach a ‘special’ grammar different from what the learner already uses in speech for him to be able to write. [3] This statement carries two important implications: that grammar (by which I think we can now understand ‘schoolroom grammar rules’) is essential for precision, and that precision is essential for writing to be effective and purposeful. No evidence is offered for either belief. Whether or not this kind of grammatical knowledge is necessary for writing to be precise is, I think questionable, although it could certainly be argued that the most difficult and sophisticated subjects require an intricate system of grammatical distinctions in order to prevent misunderstanding. It would be much harder to defend the position that without knowledge of such a system, one could not write effectively or with purpose. Popular song lyrics, for example, often contain the most flagrant deviations from the classroom norm, yet still earn their authors vast sums of money—a not inconsequential effect and presumably their intended purpose. This argument is then guilty, at the very least, of an over-generalisation fallacy, in which the word ‘writing’ is employed as a general concept, when the point made applies only to very specific forms of the activity. [4] Here, we are dealing with an empirical claim. Experience, both in work in office environments, and in dealing with various institutions, commercial and otherwise, in writing, suggests that the claim is false. There may be a grain of truth to the idea that unconventional grammar would exclude one from certain professional functions, but again, there is little reason to believe that only those who have been taught schoolroom grammar in addition to their normally acquired spoken language can be entrusted to write professionally. [5] This argument rests on the implied assumption that only ‘grammaticallycorrect’ language can be persuasive, which would be very difficult to defend. This is a case of lax expression destroying what is actually a strong argument: in the real world, it can be difficult to get ideas listened to if they are expressed in nonstandard grammar, as adherence to grammatical conventions is considered a proxy for competence, reliability and trustworthiness by many people. Rephrased, this point would be effective; but the link between standard grammar and persuasiveness is too general to be maintained.

214

12 Application of the Procedure

[6] At this point the author introduces a new topic: the connection between linguistic structure and thought. By pointing out that a poorly constructed text—even a short one like a tweet—can reflect sloppy thinking, he may be implying that the use of ‘good’ constructions has a beneficial effect on the construction of thoughts, but it is difficult to be certain. Suggesting that poor writing and poor thinking go together is one thing, suggesting that learning grammar can improve one’s thinking is something rather different. If this second is not the intention of the author, then it is unclear what relevance it has to the discussion. [7] This point is perhaps connected to the previous one, but almost certainly represents a fallacy of relevance. That criminals, who are presumably not native speakers of English, often write ungrammatically, is neither surprising nor relevant to the school curriculum in English-speaking countries. The only reason I can see to include this observation is to warn against being associated with such activities by using language containing the same type of errors. [8] It is worth noting that these last three were included in one consecutive stream of writing, as presented in the citation: for someone so committed to the cause of teaching writing skills to schoolchildren, the author certainly takes a cavalier approach to the art of paragraph construction. That aside, the point here is that grammatical knowledge is necessary for reading as well as writing. The use of the phrase ‘truly read’ is an example of both evaluatively-loaded language, and of concept manipulation: apparently, there is false reading, which any fool can do, and then there is true reading, which is better, and what reading really means. The degree to which schoolroom grammar skills are necessary for reading comprehension is debatable, particularly in the English language, which wears its grammar lightly in terms of markers for tense, case and so on. Ignorance of the subjunctive might lead one to look upon such a structure as odd, but need not result in incomprehension. [9] Similarly to point 4, this is in empirical claim that employers value workers with knowledge of grammatical conventions above those without it. No evidence is offered, and even if true this is likely to apply to a rather small range of jobs which involve producing written materials for senior staff or the public. More interestingly, the claim also implies that writing according to grammatical conventions means writing ‘properly’, and that not using those conventions leads to ambiguity, legal liability, confusion, and embarrassment. The use of ‘properly’ to describe the style he is arguing for risks creating a circularity: the proper way to write is to write properly. Of the other points, embarrassment is perhaps the most relevant, although only the most blatant examples in written materials are likely to attract much interest in the audience. As for ambiguity, legal liability and confusion, all of these can be created with perfectly standard grammar, and all of them can be avoided by the careful use of non-standard forms. This suggests that the idea of writing ‘properly’ has expanded in the author’s mind to no longer mean just writing according to grammatical convention, but to include all manner of stylistic devices and compositional techniques. The irony, given the chaotic structure, confused argument and slippery grasp of the meaning of grammar—the principle concept of the piece—present in his article, is becoming so thick as to cease to be amusing by this stage.

12.5 Example 5—A World Without Grammar

215

[10] To repeat: ‘a thinking citizen deserves the basic skills required to make sense through language, and to parse the sense and nonsense of others’. At which point irony is no longer the word; something far stronger, deeper, more unsettling, is required. The use of the verb deserve suggests a normative claim, with foundation unknown. The use of an adjective, in this case ‘thinking’, implies that those to whom the adjective does not apply are not included, that is to say that only thinking citizens deserve such skills, while unthinking citizens do not. Whence comes this notion of desert though? Are there currently thinking citizens who cannot make sense through language? If so, how do we know that they are thinking citizens at all? How have they avoided picking up the relevant skills, which are only ‘basic’ after all? The overuse of adjectives is generally considered a stylistic fault, but a fallacy of adjectives could be introduced within the clarity analysis. The two adjectives ‘thinking’ and ‘basic’ are both unnecessary and only serve to weaken the argument and confuse its expression: indeed, they come close to reducing the sentence to nonsense. If the sentence were to be rewritten as ‘a citizen requires the skills to make sense through language and understand others’ it would be rather more sensible: it would also be useful only as an argument that we should speak to our children that they might learn language. In sum then, the arguments put forward implicitly in this text, entitled It’s Not Just Grammar; It’s Clear Thinking, commit fallacies from most of the categories in the linguistic analysis assessment scheme. Most fundamentally, the author equivocates on the central term ‘grammar’, moving from meaning an essential constituent of language, to a school subject, a sanctioned and standardised codification, and finally a canon of conventions concerning the composition of written language. As well as the fallacy of equivocation, he also exhibits fallacies of definition and vagueness, as well as of relevance, and has offered unsupported and highly dubious empirical claims. That he has not thought clearly about grammar, and not just grammar, is obvious.

Conclusion

The aim of the work which led to this book was the creation of a systematic analysis procedure for assessing the linguistic features of natural language arguments. The structure of the volume, therefore, grew backwards from this final destination in order to provide a foundational background for the composition and employment of that scheme of analysis. That task has meant moving far from the usual hunting grounds of the argumentation theorist as well as looking deeply into the basic assumptions of the structures which have been built upon those grounds and constitute the various approaches, related but distinct, towards the study of arguments which are to be found in our field today. Discussion of the origin of language, and of theories of meaning, could only ever be brief, and, as a result, perhaps unsatisfying. Certain of the points made do, I believe, have bearing on how language should be considered, and, by extension, on how its role in argumentation should be studied. Perhaps the most important reason for including these brief discussions, though, was to show that they are topics of relevance to all discourse analysis: that we cannot form theories of how language is used in practice while ignoring questions of where it came from and how it comes to bear meaning. The early chapters of this book have shown that we cannot consider arguments simply as structures built out of language: we must remember that they are also built into it. Any and all expressed reasoning is expressed with words which carry argumentative content beyond that which the arguer was intending to express—or, at least, in addition to that which it was his main purpose to convey. However the first words came to be spoken, language as we know it must have developed in order to reassure, to convince, to persuade and to justify, that is, to give reasons, those reasons which were apparent to the minds of human beings and could now be shared. This same process leads, through a series of parallel steps, to the development of far greater capacity within those minds and so, just as reasoning begets argumentation through language, so language, through argumentation, begets greater reason.

© Springer Nature Switzerland AG 2021 M. Hinton, Evaluating the Language of Argument, Argumentation Library 37, https://doi.org/10.1007/978-3-030-61694-6

217

218

Conclusion

In Part II, the intention was to get a firm understanding of what arguments and argumentation are, and to survey the various ways of looking at them which have been suggested by scholars, with a particular emphasis on studies which have taken linguistic factors in argumentative discourse as their starting-point. The three chapters combine to show the richness of a field which from the outside might not be expected to demonstrate such variation and fertility. Unintentionally, these passages, I think, provide a celebration of the field of argumentation study, and an explication of the vastness of its terrain. Even for readers who cannot accept my definition of argumentation as reasoning expressed within a particular process, or who would reject the idea that there are modes of arguing which are not simply determined by the discourse situation or even the goals of the arguers, or, indeed, who would insist on a sharp division between rhetorical persuasion and logical belief formation, even for those who would reject all the elements of my approach as based on a muddled understanding of the task at hand, there is, I hope, something of interest in seeing how I have arrived at these misconceptions and how they relate to the ideas, acceptable and otherwise, of those who have put forward their own theories and observations on the phenomenon of argument and the practice of its study. The second half of the book turned from questions of what language and argument are, to questions of how they might be assessed and evaluated. The variety of ways in which arguments can go wrong has long been one of the principle objects of study in the field, and Chapter 7 could only hope to give a flavour of the existing research on fallacy theory. It is, I believe, one of the strongest points of my conception of argumentation that it leads neatly to a clear tripartite classification of fallacies. The three elements of the definition of argumentation give us three areas in which arguments can possess flaws: in their suitability to the process, in the cogency of their reasoning, and in the semantics of their expression. This typology of fallacies captures all the differing varieties of errors which have been collected under that name, leaving none unsure and adrift, while accepting happily that some argument moves may offend against good practice in more than one way and, thus, straddle the divisions. In making my definition of argumentation, and defending it, I was clear that it was a definition that would best serve my purposes and was not necessarily a prescription for others to follow. The classification of fallacies which it implies is, however, as neat, inclusive, and, most importantly, useful, as any other I have encountered, and that is a good reason why it deserves serious consideration. Part III also provided examples to illustrate the prevalence of linguistic fallacy in philosophical writing, and the degree to which philosophers have criticised the ideas of their rivals on the basis that they have misused words or misunderstood what language is capable of expressing. For anyone interested in the course of philosophical debate, this makes a clear and systematic understanding of language use in argument an urgent necessity. It also provides backing for my earlier assertion that the three types of fallacy correspond in some way to three levels of argument, with linguistic fallacies being the subtlest, most insidious, and best camouflaged, while simultaneously being most prevalent at the highest levels of intellectual endeavour, where they can have effects which shape the cultural assumptions of whole societies.

Conclusion

219

A good theory is one which leads to good practice. All of the foregoing discussion, while interesting in itself, was a preparation for the construction of the Comprehensive Assessment Procedure for Natural Argumentation presented in Chapter 11. The first two levels of the procedure, those of process and reasoning, are left as suggested frameworks to be filled out, or not, at a later time or by others. The focus here is on the third level, that of expression, which is governed by what I have termed Informal Argument Semantics. This includes considerations which have traditionally featured as linguistic fallacies, such as ambiguity and vagueness, as well as more philosophical concerns which are brought out at the ‘deep’ level of language analysis. Close study of both the tradition of fallacy theory and the complaints of philosophers about the thinking of their peers shows that arguers often go awry in their expression of reasoning by attempting to reshape the meaning of the words they use, through redefinition, or by taking the presence of that meaning to reveal certain truths about reality, rather than about its portrayal in language, which I call the fetishisation of language. In addition to this, I have shown that arguments may rely on, or produce as conclusions, statements which are semantically inconsistent. This does not occur through a fault in the logic of the argument, but through the limited possibilities for semantic cooperation contained within the words themselves. The final level of analysis is that which relies most on the theoretical considerations on the nature of language of the early chapters. A thorough investigation of the implicatures of the utterances of which the argument is formed takes in not only the work of Grice, but also the theory of argumentativity and the use of emotional and evaluative phrases in argument. It allows arguments to be fully expanded, to show their true content with nothing hidden, so that we may know with certainty just what it is that we are accepting or rejecting when we make our final judgement. The testing of the assessment scheme in Chapter 12 is not entirely fair or disinterested since I have chosen examples which show off its capabilities rather than stretch them. I am confident, however, that the structure can be applied to probe the language use of any argument, or indeed any text, to draw out the full range of semantic meaning which can be attached to it. An important feature of both the general scheme and the more detailed linguistic analysis schemes is their flexibility. There is no requirement to follow the procedure systematically from start to finish: it is possible for the experienced analyst to jump to the point at which a transgression of good practice is suspected. There are also a number of opportunities to return to earlier levels of analysis with reconstructed, fully expressed arguments, and no need to reject any argument outright where uncertainty is found. Arguments may be reformed, qualified and allowed to pass to later stages conditionally, so that even when the original argument cannot be accepted, we are left with the closest possible revision of it which can be. That an informal semantics of argument is necessary is in little doubt. While reviewing Walton’s work on ambiguity, Jim Mackenzie stated: There is reason to believe that fallacies dependent on language will become more important in intellectual life in the future. We are initiated into the discourse of one or another specialised discipline; and the discourses of different disciplines often use similar vocabulary but with tiny differences in meaning. (Mackenzie 1999: 115)

220

Conclusion

I think this is true, but understates the case. It is not only the small differences in meaning which lead us into confusion and poor argument. Technological advance has made us a world of writers, a society of arguers who seek to present our world views and persuade others on the basis of them. We are able to do this on a massive scale and with a global reach, without barriers of taste, style or coherence to block our paths to the minds of our fellow men. The examples examined in this work have shown that intellectual, serious people, even professional writers, use language wildly, unthinkingly and with a desire to confuse and manipulate. Never has it been more vital that argumentation scholars can offer a defence against this wave of incoherence and untruth; a defence which must be founded on solid theoretical fortifications, but must also take an agile, practical form in tools for assessment and analysis such as that presented in this volume.

References

Aakhus, M. (2003). Neither Naïve nor Critical Reconstruction: Dispute Mediators, Impasse, and the Design of Argumentation. Argumentation, 17, 265–290. Aberdein, A. (2006). Raising the Tone: Definition, Bullshit, and the Definition of Bullshit. In G. Reisch & G. Hardcastle (Eds.), Bullshit and Philosophy (pp. 151–169). Chicago: Open Court. Adler, J. E. (1982). Why Be Charitable? Informal Logic Newsletter, 4(2), 15–16. Almotahari, M. (2019). Semantic Deflationism Deflated. Synthese, 196(6), 2435–2454. Anscombre, J.-C. (1982). A Few Problems Relating to the Semantic Representation of Argumentation. In F. Lowenthal, F. Vandamme, & J. Cordier (Eds.), Language and Language Acquisition. Boston, MA: Springer. Anscombre, J.-C., & Ducrot, O. (1983). L’ argumentation dans la Langue. Brussels: Mardaga. Anscombre, J.-C., & Ducrot, O. (1989). Argumentativity and Informativity. In M. Meyer (Ed.), From Metaphysics to Rhetoric (pp. 71–97). Dordrecht: Kluwer Academic Publishers. Aristotle. (1955). On Sophistical Refutations (E. S. Forster, Trans.). London: William Heineman. Bach, K., & Harnish, R. (1979). Linguistic Communication and Speech Acts. Cambridge, MA: MIT Press. Balcerak Jackson, B. (2014). Verbal Disputes and Substantiveness. Erkenntnis, 79(Suppl. 1), 31–54. Baroni, P., Caminada, M., & Giacomin, M. (2011). An Introduction to Argumentation Semantics. The Knowledge Engineering Review, 26(4), 365–410. Baumtrog, M. (2018). Reasoning and Arguing, Dialectically and Dialogically, Among Individual and Multiple Participants. Argumentation, 32(1), 77–98. BBC. (1997, May 13). Jeremy Paxman & Michael Howard. Newsnight. https://www.youtube.com/ watch?v=pyqnu6ywhR4. BBC. (2014, March 23). Referee Andre Marriner Apologises for Kieran Gibbs Red Card Error. BBC Sport. https://www.bbc.com/sport/football/26699009. Belohlavek, R., & Klir, G. (Eds.). (2011). Concepts and Fuzzy Logic. Cambridge, MA: MIT Press. Bennett, B. (2018). Logically Fallacious (Academic ed.). Sudbury, MA: Archieboy Holdings. Bentham, J. ([1789] 1962). The Works of Jeremy Bentham. New York: Russell & Russell. Berkeley, G. ([1710] 1988). Principles of Human Knowledge. London: Penguin. Bilmes, J. (1991). Toward a Theory of Argument in Conversation: The Preference for Disagreement. In F. H. van Eemeren & R. Grootendorst (Eds.), Proceedings of the Second International Conference on Argumentation (Vol. 1A, pp. 462–469). Amsterdam: SICSAT. Birdsell, D., & Groarke, L. (1996). Toward a Theory of Visual Argument. Argument and Advocacy, 33, 1–10. Biro, J., & Siegel, H. (1992). Normativity, Argumentation, and an Epistemic Theory of Fallacies. In F. H. van Eemeren, R. Grootendorst, J. A. Blair, & C. A. Willard (Eds.), Argumentation Illuminated (pp. 81–103). Amsterdam: SICSAT. Bitzer, L. (1968). The Rhetorical Situation. Philosophy & Rhetoric, 1, 1–14. © Springer Nature Switzerland AG 2021 M. Hinton, Evaluating the Language of Argument, Argumentation Library 37, https://doi.org/10.1007/978-3-030-61694-6

221

222

References

Blair, J. A. (1998). The Limits of the Dialogue Model of Argument. Argumentation, 12, 325–339. Blair, J. A. (2014). Critical Review—Meta-Argumentation, an Approach to Logic and Argumentation Theory by Maurice Finocchiaro. Informal Logic, 34(2), 219–239. Bonzon, E., Delobelle, J., Konieczny, S., & Maudet, N. (2016). A Comparative Study of RankingBased Semantics for Abstract Argumentation. In Thirtieth AAAI Conference on Artificial Intelligence. https://www.aaai.org/ocs/index.php/AAAI/AAAI16/paper/viewFile/12465/11683. Brandom, R. B. (1981). Leibniz and Degrees of Perception. Journal of the History of Philosophy, 19(4), 447–449. Brandom, R. B. (2008). Between Saying and Doing: Towards an Analytic Pragmatism. Oxford: Oxford University Press. Broome, J. (2013). Rationality Through Reasoning. Chichester: Wiley Blackwell. Brown, R. (1976). Reference: In Memorial Tribute to Eric Lenneberg. Cognition, 4, 125–153. Candlish, S., & Wrisley, G. (2014). Private Language. In The Stanford Encyclopedia of Philosophy (Fall 2014 ed., E. N. Zalta, Ed.). https://plato.stanford.edu/archives/fall2014/entries/private-lan guage. Chalmers, D. (2011). Verbal Disputes. The Philosophical Review, 120(4), 515–566. Clark, H. H. (1973). The Language-as-Fixed-Effect Fallacy: A Critique of Language Statistics in Psychological Research. Journal of Verbal Learning and Verbal Behavior, 12, 335–359. Coleman, E. B. (1964). Generalizing to a Language Population. Psychological Reports, 14, 219–226. Cohen, D. (2001). Evaluating Arguments and Making Meta-Arguments. Informal Logic, 21(2), 73–84. Copi, I., Cohen, C., & McMahon, K. (2014). Introduction to Logic. Harlow: Pearson. Corredor, C. (2019). Advocacy and Enactment: Exercitives and Acts of Arguing. Research in Language, 17(1), 21–37. Darwin, C. (1888). The Expression of the Emotions in Man and Animals. New York: Appleton. Davidson, D. (1984). Inquiries into Truth and Interpretation. Oxford: Oxford University Press. Davis, W. (2014). Implicature. In Stanford Encyclopedia of Philosophy. https://plato.stanford.edu/ entries/implicature/. Demougin, D., & Fluet, C. (2006). Preponderance of Evidence. European Economic Review, 50(4), 963–976. Depperman, A. (2000). Semantic Shifts in Argumentative Processes: A Step Beyond the ‘Fallacy of Equivocation’. Argumentation, 14, 17–30. Descartes, R. (1968). Discourse on Method and the Meditations (F. E. Sutcliffe, Trans.). London: Penguin Books. Dickens, C. (1899). The Mystery of Edwin Drood. London: Chapman & Hall. Ducrot, O. (1972). Dire et ne pas Dire. Paris: Hermann. Ducrot, O. ([1983] 1983). Opérateurs Argumentatifs et Visée Argumentative. Cahiers de Linguistique Française, 5, 7–36. Dufour, M. (2017). Uncontroversial Arguments. In C. Ilie & G. Garzone (Eds.), Argumentation Across Communities of Practice: Multi-Disciplinary Perspectives (pp. 21–38). Amsterdam: John Benjamins. Dufour, M. (2017). Argument or Explanation: Who Is to Decide? Informal Logic, 37(1), 23–41. Dung, P. M. (1995). On the Acceptability of Arguments and Its Fundamental Role in Nonmonotonic Reasoning, Logic Programming and N-Person Games. Artificial Intelligence, 77(2), 321–358. Enfield, N. J. (2015). Linguistic Relativity from Reference to Agency. Annual Review of Anthropology, 44(1), 207–224. Enfield, N. J. (2015). The Utility of Meaning: What Words Mean and Why. Oxford: Oxford University Press. Engel, S. M. (1971). Wittgenstein’s Doctrine of the Tyranny of Language. The Hague: Martinus Nijhoff. Falk, D. (2004). The “Putting the Baby Down” Hypothesis: Bipedalism, Babbling, and Baby Slings. Behavioral and Brain Sciences, 27(4), 526–534.

References

223

Finocchiaro, M. (1987). Six Types of Fallaciousness: Toward a Realistic Theory of Logical Criticism. Argumentation, 1, 263–282. Finocchiaro, M. (2006). Reflections on the Hyper Dialectical Definition of Argument: Considering Pragma-Dialectics. In P. Houtlosser & A. van Rees (Eds.), Considering Pragma-Dialectics (pp. 51–62). Mahwah: Lawrence Erlbaum Associates. Finocchiaro, M. (2007). Arguments, Meta-Arguments, and Metadialogues: A Reconstruction of Krabbe, Govier, and Woods. Argumentation, 21, 253–268. Finocchiaro, M. (2013). Meta-Argumentation, an Approach to Logic and Argumentation Theory. London: College Publications. Flew, A. (1975). Thinking About Thinking (Or, Do I Sincerely Want to Be Right?). London: Fontana. Fodor, J. A. (1989). Why There Still Has to Be a Language of Thought. In P. Slezak & W. R. Albury (Eds.), Computers, Brains and Minds (Vol. 7). Australasian Studies in History and Philosophy of Science. Dordrecht: Springer. Fogelin, R. (2005). The Logic of Deep Disagreements. Informal Logic, 25(1), 3–11. Frayer, D., & Nicolay, C. (2000). Fossil Evidence for the Origin of Speech Sounds. In N. Wallin, B. Merker, & S. Brown (Eds.), The Origins of Music (pp. 217–234). Cambridge, MA: MIT Press. Frege, G. (1881). Boole’s Logical Calculus and the Concept Script. In H. Hermes, F. Kambartel, & F. Kaulbach (Eds.), Posthumous Writings (P. Long & R. White, Trans., pp. 9–46). Chicago: University of Chicago Press. Frege, G. ([1964] 1882). On the Scientific Justification of a Concept-Script (J. M. Bartlett, Trans.). Mind, LXXIII(290), 155–160. Gentleman, A. (2018, February 23). Northern Irish Abortion Law Violates Women’s Rights, Say UN Officials. The Guardian. https://www.theguardian.com/uk-news/2018/feb/23/northern-irelandabortion-law-violates-womens-rights-says-un-committee. Gentner, D., Bowdle, B., Wolff, P., & Boronat, C. (2001). Metaphor Is Like Analogy. In D. Gentner, K. J. Holyoak, & B. N. Kokinov (Eds.), The Analogical Mind: Perspectives from Cognitive Science (pp. 199–253). Cambridge, MA: MIT Press. Gettier, E. (1963). Is Justified True Belief Knowledge? Analysis, 23(6), 121–123. Gilbert, M. (2002). Effing the Ineffable: The Logocentric Fallacy in Argumentation. Argumentation, 16, 21–32. Gilbert, M. (2004). Emotion, Argumentation and Informal Logic. Informal Logic, 24(3), 245–264. Goodwin, J. (2007). Argument Has No Function. Informal Logic, 27(1), 69–90. Goodwin, J., & Innocenti, B. (2016). The Pragmatic Force of Making Reasons Apparent. In D. Mohammed & M. Lewi´nski (Eds.), Argumentation and Reasoned Action: Proceedings of the 1st European Conference on Argumentation, Lisbon, 2015 (Vol. II, pp. 449–462). London: College Publications. Govier, T. (1981). Uncharitable Thoughts About Charity. Informal Logic Newsletter, 4(20), 5–6. Govier, T. (1982). On Adler on Charity. Informal Logic Newsletter, 4(4), 10–11. Govier, T. (2010). A Practical Study of Argument (7th ed.). Belmont, CA: Wadsworth. Graham, A. C. (1989). Disputers of the Tao: Philosophical Argument in Ancient China. Chicago and La Salle: Open Court. Grandin, T. (2006). Thinking in Pictures. New York: Vintage Press. Green, M. (2010). Speech Acts. In T. O’Connor & C. Sandis (Eds.), A Companion to the Philosophy of Action (pp. 58–66). Oxford: Wiley-Blackwell. 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. (1989). Studies in the Way of Words. Cambridge, MA: Harvard University Press. Grize, J.-B. (1986). Raisonner en Parlant [Reasoning While Speaking]. In M. Meyer (Ed.), De la Métaphysique à la Rhétorique Essais à la Mémoire de Chaïm Perelman avec un Inédit sur la Logique (pp. 45–55). Bruxelles: Editions de l’Université de Bruxelles. Hamblin, C. L. (1970). Fallacies. London: Methuen. Hample, D., & Irions, A. (2015). Arguing to Display Identity. Argumentation, 29(4), 381–416.

224

References

Hansen, H. (2002). The Straw Thing of Fallacy Theory: The Standard Definition of Fallacy. Argumentation, 16, 133–155. Habernal, I., Wachsmuth, H., Gurevych, I., & Stein, B. (2018). The Argument Reasoning Comprehension Task: Identification and Reconstruction of Implicit Warrants. In Proceedings of NAACL-HLT 2018 (pp. 1930–1940). https://www.aclweb.org/anthology/N18-1175. Hazen, M., & Williams, D. (1997). Argument and Identity. Argument and Advocacy, 34, 5–8. Hinton, M. (2016). Identity as Argumentation: Argumentation as Identity. In K. Ciepiela (Ed.), Identity in Communicative Contexts (Vol. 48, pp 177–188). Łód˙z Studies in Language. Frankfurt am Main: Peter Lang. Hinton, M. (2020a). The Platform Fallacy—A Dickensian Contribution to Informal Logic. Philosophy & Literature, 44(2), 449-460. Hinton, M. (2020b). Extracting Meta-Arguments from Online Corpora. Argumentation. https://doi. org/10.1007/s10503-020-09533-z. Hinton, M., & Budzy´nska-Daca, A. (2019). A Comparative Study of Political Communication in Televised Pre-Election Debates in Poland and the United States of America. Research in Language, 17(1), 1–19. Hinton, M., & Wagemans, J. (forthcoming). Evaluating the Underlying Reasoning of Arguments: A Procedural Approach. Hirsch, E. (2005). Physical-Object Ontology, Verbal Disputes, and Common Sense. Philosophy and Phenomenological Research, 70, 67–97. Howarth, S, & Mundy, J. (2015). Marcel Duchamp—Fountain. Tate.org. https://www.tate.org.uk/ art/artworks/duchamp-fountain-t07573. Huppi, T. (n.d.). Myth: There’s No Such Thing as Society… Only Individuals and Families. http:// www.huppi.com/kangaroo/L-nosociety.htm. Hurley, P. J. (2000). A Concise Introduction to Logic (7th ed.). Belmont, CA: Wadsworth. IFAB. (2019a). Video Assistant Referees (VARs) Used Live in Competitions and Leagues. Theifab.com. http://theifab.com/projects/vars/background-scope. IFAB. (2019b). Laws of the Game 2019/20. IFAB. http://static-3eb8.kxcdn.com/documents/793/ 103202_200519_LotG_201920_EN_SinglePage.pdf. Ipsos MORI. (2013). No Such Thing as Society?https://www.ipsos.com/ipsos-mori/en-uk/no-suchthing-society. Jackson, S. (2015). Design Thinking in Argumentation Theory and Practice. Argumentation, 29, 243–263. Jackson, S., & Jacobs, S. (2006). Designing Countermoves to Questionable Argumentative Tactics. In F. H. van Eemeren, M. D. Hazen, P. Houtlosser, & D. C. Williams (Eds.), Contemporary Perspectives on Argumentation: Views from the Venice Argumentation Conference (pp. 83–100). Amsterdam: International Centre for the Study of Argumentation. Jacob, P. (2019). Intentionality. The Stanford Encyclopedia of Philosophy (Spring 2019 ed., E. N. Zalta, Ed.). https://plato.stanford.edu/archives/spr2019/entries/intentionality/. Jacobs, S. (2006). Nonfallacious Rhetorical Strategies: Lyndon Johnson’s Daisy Ad. Argumentation, 20, 421–442. Jason, G. (1988). Hedging as a Fallacy of Language. Informal Logic, X(3), 169–175. Javez, J., & Tovena, M. (2008). Presque and Almost: How Argumentation Derives from Comparative Meaning. In O. Bonami & P. Cabredo Hofherr (Eds.), Empirical Issues in Syntax and Semantics 7 (pp. 217–239). Paris: Colloque de syntaxe et sémantique à Paris. Jenkins, C. S. I. (2014). Merely Verbal Disputes. Erkenntnis, 79, 11–30. Jenkins, S. (2019, May 31). What Are Our Universities for? Taxpayers Have a Right to Know. The Guardian. https://www.theguardian.com/commentisfree/2019/may/31/what-are-uni versities-for-taxpayers-right-know. Johnson, R. H. (1981). Charity Begins at Home. Informal Logic Newsletter, 3(3), 4–9. Johnson, R. H. (2000). Manifest Rationality: A Pragmatic Theory of Argument. Mahwah, NJ: Lawrence Erlbaum Associates. Johnson, R. H., & Blair, J. A. (2006). Logical Self-Defense. New York: Idebate Press.

References

225

Joule, B. (2015). Lost Art: I Destroyed a Bacon Masterpiece. Tate. https://www.tate.org.uk/contextcomment/video/lost-art-i-destroyed-bacon-masterpiece. Kaplan, D. (1989). Afterthoughts. In J. Almog, J. Perry, & H. Wettstein (Eds.), Themes from Kaplan (pp. 565–614). Oxford: Oxford University Press. Kay, P., & Kempton, W. (1984). What Is the Sapir-Whorf Hypothesis? American Anthropologist, 86(1), 65–79. Kienpointner, M. (1996). Whorf and Wittgenstein: Language, World View and Argumentation. Argumentation, 10, 475–494. Kirwan, C. (1979). Aristotle and the So-Called Fallacy of Equivocation. The Philosophical Quarterly, 29(114), 35–46. Kjeldsen, J. (2018). The Rhetoric of Sound, the Sound of Arguments Three Propositions, Three Questions, and an Afterthought for the Study of Sonic and Multimodal Argumentation. Argumentation and Advocacy, 54(4), 347–354. Kneale, W., & Kneale, M. (1962). The Development of Logic. Oxford: Clarendon Press. Kockelman, P. (2013). Agent, Person, Subject, Self: A Theory of Ontology, Interaction, and Infrastructure. Oxford: Oxford University Press. Krabbe, E. (1998). Who Is Afraid of Figure of Speech? Argumentation, 12, 281–294. Kup´s, P. (2019). Knowledge Representation Systems and Natural Language Argumentation. Unpublished B.A. thesis, University of Łód´z. Labour Party. (2017). For the Many, Not the Few. The Labour Party Manifesto 2017. http://labour. org.uk/wp-content/uploads/2017/10/labour-manifesto-2017.pdf. Lakatos, I. (1976). Proofs and Refutations: The Logic of Mathematical Discovery (J. Worrall & E. Zahar, Eds.). Cambridge: Cambridge University Press. Lakoff, G. (1973). Hedges: A Study in Meaning Criteria and the Logic of Fuzzy Concepts. Journal of Philosophical Logic, 2, 458–508. Lakoff, G. (1987). Women, Fire and Dangerous Things. Chicago: University of Chicago Press. Lao Tzu. (1963). Tao Te Ching (D. C. Lau, Trans.). London: Penguin. Leech, G. (1983). Principles of Pragmatics. London: Longman. Leszczy´nksa-Jasion, D., & Łupkowski, P. (2016). Erotetic Search Scenarios and Three-Valued Logic. Journal of Logic, Language and Information, 25(1), 55–76. Levinson, S. (1983). Pragmatics. Cambridge: Cambridge University Press. Lewi´nski, M. (2012). The Paradox of Charity. Informal Logic, 32(4), 403–439. Locke, J. ([1690] 1975). An Essay Concerning Human Understanding. Oxford: Oxford University Press. Lucy, J. (1997). Linguistic Relativity. Annual Review of Anthropology, 26, 291–312. Łupkowski, P. (2018). Erotetic Reasoning Corpus: A Data Set for Research on Natural Question Processing. Journal of Language Modelling, 5, 607–631. Macagno, F. (2012). Presumptive Reasoning in Interpretation: Implicatures and Conflicts of Presumptions. Argumentation, 26, 233–265. Macagno, F. (2014). Manipulating Emotions: Value-Based Reasoning and Emotive Language. Argumentation and Advocacy, 51, 103–122. Macagno, F., & Walton, D. (2014). Implicatures as Forms of Argument. In A. Capone, et al. (Eds.), Perspectives on Pragmatics and Philosophy (pp. 203–224). Berlin and New York: Springer. Macagno, F., & Zavatta, B. (2014). Reconstructing Metaphorical Meaning. Argumentation, 28, 453–488. Mackenzie, J. (1999). Book Review: Walton, Douglas (Ed.), Fallacies Arising from Ambiguity (1996). Argumentation, 13, 115–119. Mackenzie, J. (2014). From Speech Acts to Semantics. Studies in Logic, Grammar and Rhetoric, 36(49), 121–142. Mahdawi, A. (2019, August 28). How Often Do Americans Change Their Underwear? The Statistics Stink. The Guardian. https://www.theguardian.com/commentisfree/2019/aug/21/how-often-doamericans-change-their-underwear-the-statistics-stink#comments.

226

References

Martini, C. (2018). Ad Hominem Arguments, Rhetoric, and Science Communication. Studies in Logic, Grammar & Rhetoric, 55(68), 151–166. Mauthner, F. (1901–1902). Beiträge zu einer Kritik der Sprache. Stuttgart and Berlin: Cotta. McBurney, P., & Parsons, S. (2001). Chance Discovery Using Dialectical Argumentation. In T. Terano, T. Nishida, A. Namatame, S. Tsumoto, Y. Ohsawa, & T. Washio (Eds.), New Frontiers in Artificial Intelligence (Vol. 2253, pp. 414–424). Lecture Notes in Artificial Intelligence. Berlin: Springer Verlag. Metcalf, R. (2005). Rethinking the Ad Hominem: A Case Study of Chomsky. Argumentation, 19, 29–52. Mill, J. S. (1882). A System of Logic, Ratiocinative and Inductive (8th ed.). New York: Harper & Brothers. Monty Python. (1979). Life of Brian. http://montypython.50webs.com/scripts/Life_of_Brian/8.htm. Moore, G. E. (1903). Principia Ethica. London: Cambridge University Press. Müller, M. (1864). Lectures on the Science of Language: Delivered at the Royal Institution of Great Britain in April, May, & June, 1861, Volume 1: The Science of Language. London: Longman, Green, Longman, Roberts & Green. Naess, A. (1966). Communication and Argument. London: Allen and Unwin. Nájera, E. (2007). Wittgenstein Versus Mauthner: Two Critiques of Language, Two Mysticisms. In H. Hrachovec, A. Pichler, & J. Wang (Eds.), Papers of the 30th International Wittgenstein Symposia (pp. 160–162). Kirchberg am Wechsel: ALWS. Nelson, L. (2016). A Theory of Philosophical Fallacies (F. Leal & D. Carus, Trans.). Cham: Springer. Office for National Statistics. (2017). Persistent Poverty in the UK and EU: 2015. https://www. ons.gov.uk/peoplepopulationandcommunity/personalandhouseholdfinances/incomeandwealth/ articles/persistentpovertyintheukandeu/2015. Ogden, M. (2009, November 19). Thierry Henry Admits to Handball That Defeated Ireland in World Cup Play-Off. The Telegraph. https://www.telegraph.co.uk/sport/football/teams/rep ublic-of-ireland/6599687/Thierry-Henry-admits-to-handball-that-defeated-Ireland-in-WorldCup-play-off.html. Okamoto, S. (1993). Nominal Repetitive Constructions in Japanese: The ‘Tautology’ Controversy Revisited. Journal of Pragmatics, 20(5), 433–466. Okundia, J., & Efunla, A. (2019). Substance Abuse: Lizzy Should Have Been Rehabilitated First— Expert. PM News. https://www.pmnewsnigeria.com/2019/04/30/substance-abuse-lizzy-shouldhave-been-rehabilitated-first-expert/. Oswald, S., Herman, T., & Jacquin, J. (Eds.). (2018). Argumentation and Language—Linguistic, Cognitive and Discursive Explorations. Cham: Springer. Oswald, S., & Rihs, A. (2014). Metaphor as Argument: Rhetorical and Epistemic Advantages of Extended Metaphors. Argumentation, 28, 133–159. Parson D. W., & Ziegelmueller, G. (2003). Linguistically Sound Arguments: Part II. Eloquence and Argument. In F. H. van Eemeren, J. A. Blair, C. A. Willard, & A. F. Snoeck-Henkemans (Eds.), Proceedings of the Fifth Conference of the International Society for the Study of Argumentation (pp. 811–814). Amsterdam: Sic Sat. Partington, A. (2003). The Linguistics of Political Argument. London: Routledge. Partington, R. (2018). Four Million British Workers Live in Poverty, Charity Says. The Guardian. https://www.theguardian.com/business/2018/dec/04/four-million-british-workers-live-in-pov erty-charity-says. Patterson, S. (2011). Functionalism, Normativity and the Concept of Argumentation. Informal Logic, 31(1), 1–26. Pears, D. (1988). The False Prison—A Study of the Development of Wittgenstein’s Philosophy (Vol. 2). Oxford: Oxford University Press. Penn, J. (1972). Linguistic Relativity Versus Innate Ideas: The Origins of the Sapir-Whorf Hypothesis in German Thought. Paris: Mouton. Perelman, C., & Olbrechts-Tyteca, L. (1969). The New Rhetoric. Notre Dame: University of Notre Dame Press.

References

227

Pérez, D. (2005). Is Thought Without Language Possible? Principia, 9(1–2), 177–191. Plantin, C. (2009). A Place for Figures of Speech in Argumentation Theory. Argumentation, 23, 325–337. Poliva, O. (2015). From Where to What: A Neuroanatomically Based Evolutionary Model of the Emergence of Speech in Humans. F1000Research, 4, 67. Poliva, O. (2016). From Mimicry to Language: A Neuroanatomically Based Evolutionary Model of the Emergence of Vocal Language. Frontiers in Psychology, 10(307), 1–21. Raaijmakers, J., Schrijnemakers, J., & Gremmen, F. (1999). How to Deal with “The Language-asFixed-Effect Fallacy”: Common Misconceptions and Alternative Solutions. Journal of Memory and Language, 41, 416–426. Rawls, J. (1972). A Theory of Justice. Oxford: Oxford University Press. Reboul, A. (1989). Relevance and Argumentation: How Bald Can You Get. Argumentation, 3, 285–302. Reps, P. (1971). Zen Flesh, Zen Bones. Harmondsworth: Penguin. Rott, H. (2015). A Puzzle About Disputes and Disagreements. Erkenntnis, 80, 167–189. Rushkoff, D. (2013, January 30). It’s Not Just Grammar; It’s Clear Thinking. The New York Times. https://www.nytimes.com/roomfordebate/2012/08/13/is-our-children-learning-eno ugh-grammar-to-get-hired/its-not-just-grammar-its-clear-thinking. Scholz, B., Pelletier, F., & Pullum, G. (2016). Philosophy of Linguistics. In E. N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy (Winter 2016 ed.), https://plato.stanford.edu/archives/win 2016/entries/linguistics. Schreiber, S. (2003). Aristotle on False Reasoning. Albany: State University of New York Press. Scriven, M. (1976). Reasoning. New York: McGraw-Hill. Searle, J. (1979). Expression and Meaning: Studies in the Theory of Speech Acts. Cambridge: Cambridge University Press. Shapiro, B. (1991). “Beyond Reasonable Doubt” and “Probable cause”: Perspectives on the AngloAmerican Law of Evidence. Berkeley: University of California Press. Sharma, J. (2019). Thousands of Foreign Students Caught in Limbo by Scandal. University World News. https://www.universityworldnews.com/post.php?story=20190207192224757. Sider, T. (2009). Ontological Realism. In D. Chalmers, D. Manley, & R. Wasserman (Eds.), Metametaphysics: New Essays on the Foundations of Ontology. Oxford: Oxford University Press. Siegel, H., & Biro, J. (1997). Epistemic Normativity, Argumentation and Fallacies. Argumentation, 11, 277–292. Slade, C. (1995). Reflective Reasoning in Groups. Informal Logic, 17(2), 223–234. Smith, B. (2006). Why Polish Philosophy Does Not Exist. In J. J. Jadacki & J. Pasniczek (Eds.), The Lvov-Warsaw School: The New Generation (Vol. 89, pp. 19–39). Poznan Studies in the Philosophy of the Sciences and the Humanities. Amsterdam: Rodopi. Snoeck Henkemans, A. F. (1992). Analysing Complex Argumentation. Amsterdam: SicSat. Snoeck Henkemans, A. F. (2014). Speech Act Theory and the Study of Argumentation. Studies in Logic, Grammar and Rhetoric, 36(49), 41–58. Sorensen, R. (2001). Vagueness and Contradiction. Oxford: Oxford University Press. Sorensen, R. (2018). Vagueness. The Stanford Encyclopedia of Philosophy (Summer 2018 ed., E. Zalta, Ed.). https://plato.stanford.edu/archives/sum2018/entries/vagueness. Spengler, O. ([1918] 1991). The Decline of the West (C. F. Atkinson, Trans.). Oxford: Oxford University Press. Sperber, D., & Wilson, D. (1986). Loose Talk. Proceedings of the Aristotelian Society, 1985–6, 153–171. Stark, H. (2000). The Lord Scroop Fallacy. Informal Logic, 20(3), 245–259. Stevenson, C. (1944). Ethics and Language. New Haven, CT: Yale University Press. Thatcher, M. (1987). Interview for Woman’s Own (“No Such Thing as Society”). https://www.mar garetthatcher.org/document/106689. Tindale, C. (1996). Fallacies in Transition: An Assessment of the Pragma-Dialectical Perspective. Informal Logic, 18(1), 17–33.

228

References

Tindale, C. (2007). Fallacies and Argument Appraisal. Cambridge: Cambridge University Press. Toulmin, S. ([1958] 2003). The Use of Argument. Cambridge: Cambridge University Press. Truss, L. (2003). Eats, Shoots and Leaves. London: Profile Books. Urba´nski, M., van Lambalgen, M., & Koszowy, M. (2017). An Introduction to the Special Issue on Question Processing. Logic & Logical Philosophy, 26(3), 285–288. van Eemeren, F. H. (1986). Dialectical Analysis as a Normative Reconstruction of Argumentative Discourse. Text: An Interdisciplinary Journal for the Study of Discourse, 6(1), 1–16. van Eemeren, F. H. (2010). Strategic Maneuvering in Argumentative Discourse: Extending the Pragma-Dialectical Theory of Argumentation. Amsterdam and Philadelphia: John Benjamins. van Eemeren, F. H. (2015). Reasonableness and Effectiveness in Argumentative Discourse. Argumentation Library 27. Cham: Springer. van Eemeren, F. H., Garssen, B. J., Krabbe, E. C. W., Snoeck Henkemans, A. F., Verheij, B., & Wagemans, J. H. M. (2014). Handbook of Argumentation Theory. Dordrecht: Springer. van Eemeren, F. H., & Grootendorst, R. (1992). Argumentation, Communication, and Fallacies: A Pragma-Dialectical Perspective. Hillsdale, NJ: Lawrence Erlbaum Associates, Publishers. van Eemeren, F. H., & Grootendorst, R. (1992). Relevance Reviewed: The Case of Argumentum ad Hominem. Argumentation, 6, 141–159. van Eemeren, F. H., & Grootendorst, R. (2004). A Systematic Theory of Argumentation. Cambridge: Cambridge University Press. van Eemeren, F. H., & Houtlosser, P. (2005, May 18–21). Theoretical Construction and Argumentative Reality: An Analytic Model of Critical Discussion and Conventionalised Types of Argumentative Activity. In D. Hitchcock (Ed.), The Uses of Argument: Proceedings of a Conference at McMaster University. Ontario Society for the Study of Argumentation. Windsor, ON: OSSA. van Eemeren, F. H., & Houtlosser, P. (2015). Strategic Maneuvering: Examining Argumentation in Context. In F. H. van Eemeren (Ed.), Reasonableness and Effectiveness in Argumentative Discourse. Cham: Springer. van Eemeren, F. H., & Houtlosser, P. (2015). The Study of Argumentation as Normative Pragmatics. In F. H. van Eemeren (Ed.), Reasonableness and Effectiveness in Argumentative Discourse. Cham: Springer. van Laar, J. A. (2003). The Dialectic of Ambiguity. Groningen: University of Groningen. Vatz, R. (1973). The Myth of the Rhetorical Situation. Philosophy & Rhetoric, 6(3), 154–161. Vermeulen, I. (2018). Verbal Disputes and the Varieties of Verbalness. Erkenntnis, 83, 331–348. Visser, J., Koszowy, M., Konat, B., Budzy´nska, K., & Reed, C. (2018). Straw Man as Misuse of Rephrase. In S. Oswald & D. Maillat (Eds.), Argumentation and Inference: Proceedings of the 2nd European Conference on Argumentation, Fribourg 2017 (Vol. 2). London: College Publications. Wagemans, J. H. M. (2016, May 18–21). 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, ON: OSSA. Wagemans, J. H. M. (2019). Four Basic Argument Forms. Research in Language, 17(1), 57–69. Walton, D. (1993). Commitment, Types of Dialogue, and Fallacies. Informal Logic, 14(2&3), 93– 103. Walton, D. (1997). Appeal to Expert Opinion: Arguments from Authority. University Park, PA: Penn State University Press. Walton, D. (2004). Classification of Fallacies of Relevance. Informal Logic, 24(1), 71–103. Walton, D. (2010). Types of Dialogue and Burdens of Proof. In P. Baroni, F. Cerutti, M. Giacomin, & G. Simari (Eds.), Computational Models of Argument. Amsterdam: IOS Press BV. Walton, D. (2020). Syntactic Parsers and Profiles of Dialogue Applied to the Fallacy of Amphiboly. Informal Logic, 40(1), 3–45. Walton, D., & Krabbe, E. (1995). Commitment in Dialogue. Albany: State University of New York Press.

References

229

Walton, D., Reed, C., & Macagno, F. (2008). Argumentation Schemes. Cambridge: Cambridge University Press. Weiler, G. (1958). On Fritz Mauthner’s Critique of Language. Mind, 67(265), 80–87. Whatley, R. (1857). Elements of Logic (9th ed.). London: John Parker and Son. Whorf, B. (1956). Language, Thought and Reality. Cambridge, MA: MIT Press. Wildgen, W. (2004). The Evolution of Human Language. Amsterdam: John Benjamins. Wilkinson, M. (2019). The 50 Weirdest Job Titles. Coburg Banks. https://www.coburgbanks.co.uk/ blog/friday-funnies/the-50-weirdest-job-titles/. Wilson, D., & Sperber, D. (1986). Inference and Implicature in Utterance Interpretation. In T. Myers, K. Brown, & B. McGonigle (Eds.), Reasoning and Discourse Processes (pp. 241–263). London: Academic Press. Wittgenstein, L. (1953). Philosophical Investigations (G. E. M. Anscombe, Trans.). Oxford: Blackwell. Wittgenstein, L. (1958). The Blue and Brown Books, Preliminary Studies for the Philosophical Investigations. London: Blackwell. Wittgenstein, L. (1974). Tractatus Logico-Philosophicus (D. F. Pears & B. F. McGuinness, Trans.). London: Routledge. Wohlrapp, H. (2014). The Concept of Argument. Dordrecht: Springer. Wolff, P., & Holmes, K. (2011). Linguistic Relativity. Wiley Interdisciplinary Review of Cognitive Science, 2(3), 253–265. Woods, J. (1988). Pragma-Dialectics: A Radical Departure in Fallacy Theory. ISSA Newsletter, 4, 5–15. Woods, J. (2007). The Concept of Fallacy Is Empty. In L. Magnani & P. Li (Eds.), Model-Based Reasoning in Science, Technology, and Medicine: Studies in Computational Intelligence (Vol. 64, pp. 69–90). Berlin: Springer. Woods, J., & Walton, D. (1975). Petitio Principii. Synthese, 31, 107–127. Yu, S., & Zenker, F. (2020). Schemes, Critical Questions, and Complete Argument Evaluation. Argumentation, 34, 469–498. Zadeh, L. A. (1965). Fuzzy Sets. Information and Control, 8, 338–353. Zenker, F. (Ed.). (2013). Bayesian Argumentation—The Practical Side of Probability. Dordrecht: Springer. Zlatev, J., & Blomberg, J. (2015). Language May Indeed Influence Thought. Frontiers in Psychology, 6(1631), 1–10.

Index

A Aakhus, Mark, 158 Aberdein, Andrew, 130 abstract, 20, 113, 137, 189 abstract concepts, 20, 142 ad hominem, 103, 105 ad verecundiam, 103, 105 ambiguity, 15, 20, 73, 86, 91, 102, 109, 117, 118, 129, 140, 146, 158, 163, 182, 187, 193, 198 ambiguous language, 21 amphiboly, 110, 113, 115, 183, 193 analogy, 86 analysis, 3, 56, 65, 71, 85, 107, 167 ambiguity analysis, 183 concept analysis, 189, 193 deep linguistic analysis, 163, 186, 187, 193 implicature analysis, 191, 200 initial analysis, 158, 168, 195 linguistic analysis, 85, 91, 145, 178 process analysis, 172 reasoning analysis, 85, 165, 174 verbal analysis, 89 Anscombre, Jean-Claude, 31, 38, 79 appropriateness, 172 argument, 3, 47, 51, 52, 55, 65, 67, 105, 173 argument and argumentation, 51 argument from ignorance, 100 argument reconstruction, 82, 91, 154, 157, 159, 168, 171, 184, 202 argument reconstruction guidelines, 160 argument reformulation, 182, 191 argument strength, 76, 162, 177 argument type identification procedure, 170

complex arguments, 165 condensed arguments, 125 emotional arguments, 124 first-order arguments, 174, 197 predicate arguments, 174, 197 second-order arguments, 174 subject arguments, 174 argumentation, 45, 51, 56–58, 64, 67, 81, 88, 89, 91, 137, 149, 161 antagonistic argumentation, 71 argumentation, definition of, 48 argumentation frameworks, 178 argumentation, modes of, 48, 68, 172, 202 argumentation, theory of, 8, 13, 80, 89, 91, 111, 156 collaborative argumentation, 63 compound argumentation, 165 cooperative argumentation, 46, 62, 65, 67, 70 demonstrative argumentation, 76 formal argumentation, 149 monological argumentation, 45, 69, 70, 157 visual argumentation, 48, 50 argumentativity, 38, 42, 124, 125, 160, 191, 209 argument schemes, 52, 91, 103, 104, 165, 174, 176 Aristotle, 86, 98, 100, 107, 109, 123, 126, 136 assessment, 167

B Baumtrog, Michael, 45, 69, 70 begging the question, 125, 130, 132

© Springer Nature Switzerland AG 2021 M. Hinton, Evaluating the Language of Argument, Argumentation Library 37, https://doi.org/10.1007/978-3-030-61694-6

231

232 belief, 7, 47, 77, 104, 136, 161, 187 Bentham, Jeremy, 139, 148 Berkeley, George, 113, 142, 144, 147, 189 Bitzer, Lloyd, 87 Blair, J. Anthony, 102, 119, 124, 127, 157, 164, 166 Brandom, Robert, 84

C CAPNA - comprehensive assessment procedure for natural argumentation, 167, 195, 219 Chalmers, David, 89 clarity, 84, 122, 127, 138, 148, 179, 180, 183, 187 coding, 47, 48 cognitive science, 82 coherence, 8, 60, 163, 179, 187, 193 completeness, 180 concepts, 7, 11, 21, 84, 89, 91, 107, 138, 141, 163, 180, 189, 214 fuzzy concepts, 120 conceptualisation, 3, 10, 12, 15, 86, 180 conflict, 46, 62, 63, 65–67, 71, 77, 111 consistency, 179 conversational maxims, 34 critical discussion, 26, 57, 62–64, 72, 83, 98, 159, 161, 164 critical questions, 37, 55, 91, 117, 165, 174

D Davidson, Donald, 7, 159 definition, 21, 46, 48, 58, 90, 91, 138, 139, 187, 193, 197, 199 persuasive definition, 130, 139, 140, 189, 200 Descartes, René, 6 dialectic, 59, 65, 74, 85, 138 Dickens, Charles, 103 disagreement, 46, 49, 65, 69, 71, 89, 92 deep disagreement, 14, 90, 190 discourse, 6, 80, 81, 83, 86–88, 92, 100, 111, 154, 162, 171, 175 Ducrot, Oswald, 31, 38, 79, 88 Dufour, Michel, 65, 69, 77

E equivocation, 28, 61, 73, 86, 92, 102, 107, 110, 117, 130, 131, 180, 184, 193, 208, 211, 213 synonymic equivocation, 185, 206

Index evaluation, 66, 82, 159, 163, 167, 175 explanation, 6, 65 expression, 47, 70, 83, 105 expression of reasoning, 48 F fallacies, 27, 58, 60, 81, 87, 97, 104, 153 fallacies of coherence, 189 fallacies of concept, 189 fallacies of definition, 189 fallacies of expression, 61, 70, 72, 106, 107, 109, 114, 167, 179 fallacies of language, 50, 70, 73, 86, 91, 101, 109, 116, 129, 135, 148 fallacies of process, 61, 72, 105, 107, 122, 172 fallacies of reasoning, 61, 106, 107 fallacies of writing, 116 fallacies, typology of, 100, 109, 154 fallaciousness, 102 fallacy of accent, 115 fallacy of adjectives, 215 fallacy of composition, 115, 203, 205 fallacy of division, 115 fallacy, standard definition of, 98 fallacy theory, 91, 105 formal fallacies, 104 gambler’s fallacy, 104 generalisation fallacy, 104, 213 informal fallacies, 104 language-as-fixed-effect fallacy, 104 logical fallacies, 98, 106 Lord Scroop fallacy. See self-defeating statements naturalistic fallacy, 90, 139, 141, 148, 189 no-true-Scotsman fallacy, 130 philosophical fallacy, 99, 107, 139, 148, 189 platform fallacy, 103 straw man fallacy, 48, 92, 103, 105, 118, 172, 203, 206 fetishisation of language, 66, 147, 189 figures, 86–88, 113 figurative language, 132, 208 Finocchiaro, Maurice, 65, 102, 156 Fogelin, Robert, 14, 90, 190 Frege, Gottlob, 8, 84 G gerrymandering a term, 130, 131, 189 Gilbert, Michael, 48, 124

Index Goodwin, Jean, 67, 76 Govier, Trudy, 159, 165 Grice, Paul, 31, 32, 34, 35, 83, 91, 157, 219 Grize, Jean-Blaise, 80 Grootendorst, Rob, 42, 57, 69, 82, 85, 98, 102, 117, 161

H Hamblin, Charles, 84, 98, 114 Hansen, Hans, 98 hedging, 128 homonymy, 109, 113

I implicature, 31, 32, 35, 42, 91, 159, 163, 171, 180, 191, 193 inferentialism, 84 informal logic, 50, 83, 160

J Jackson, Sally, 167 Jacobs, Scott, 87, 167 Johnson, Ralph, 57, 102, 119, 124, 127, 159, 166 justification, 66

K Kant, Immanuel, 49, 84, 138 Kienpointner, Manfred, 13, 127 knowledge, 27, 47, 140, 144, 161 Krabbe, Eric, 62, 114

L Lakoff, George, 13, 121 language, 3, 5, 6, 24, 47, 89, 133, 136, 143, 144, 186 emotive language, 77, 124, 191, 200 evaluative language, 77, 124, 191, 200 ideal language, 144 language, critique of, 143, 144 language use, principles of, 180 loaded language, 75, 123, 131, 214 language games, 13, 25 Lao Tzu, 137, 144 levers, 162, 175, 176 Lewi´nski, Marcin, 159 linguistic relativity, 9, 13 linguistics, 4, 79, 82, 89, 91, 178, 191 Locke, John, 87, 105, 137, 142

233 logic, 29, 59, 61, 80, 138, 146 description logics, 179 formal logic, 80, 127, 161 natural logic, 80

M Macagno, Fabrizio, 35, 36, 86, 124, 127 Mackenzie, Jim, 84, 219 Mauthner, Fritz, 143, 147, 189 meaning, 17, 21, 39, 84, 89, 91, 121, 130, 135, 146, 178, 211 meta-argument, 52, 71, 154, 156, 173 metaphor, 14, 86 Mill, John Stuart, 101, 114 Moore, G.E., 90, 139, 147, 189 mysticism, 15, 137, 145, 147, 190

N Naess, Arne, 11, 58, 72, 102, 120, 124, 140, 141 Nelson, Leonard, 99, 107, 137, 140, 147, 189 normativity, 67, 82, 97, 99, 155, 160

O Olbrechts-Tyteca, Lucie, 58, 88, 130 Oswald, Steve, 79, 81, 86

P paronymous words, 114, 141, 189 Patterson, Steven, 67, 69 Pears, David, 25 Perelman, Chaïm, 88, 130 periodic table of arguments, 170, 174, 176, 185, 197 persuasion, 6, 61, 72, 73, 86, 124, 161, 213 philosophy, 19, 21, 24, 49, 87, 89, 91, 107, 135, 149, 191 continental philosophy, 123, 148 philosophy of language, 17, 27, 89, 147 Plantin, Christian, 87, 88 Plato, 74, 114, 141, 142 pragma-dialectical approach, 15, 26, 57, 61, 63, 65, 69, 82, 98, 121, 159, 161, 163, 172, 178 pragmatics, 28, 66, 68, 79, 84, 162, 178 informal argument pragmatics, 172 normative pragmatics, 161 principle of charity, 159 procedural questions, 172, 176, 193 process, 48, 57, 58, 65, 70, 83, 105, 162, 172

234 R radical interpretation, 8, 159 rationale, 65, 77 rationality, 8, 57, 58, 85, 161, 162 Rawls, John, 149 reasonableness, 82, 161 reasoning, 3, 24, 45, 47, 51, 57, 65, 70, 76, 80, 81, 83, 85, 91, 100, 101, 105, 115, 129, 137, 162, 170 abductive reasoning, 12, 54 circular reasoning, 105. See also begging the question deductive reasoning, 25, 54, 126 defeasible reasoning, 12 emotional reasoning, 125 erotetic reasoning, 74 inductive reasoning, 54 practical reasoning, 12 presumptive reasoning, 54, 104 underlying reasoning, 176 reification, 86, 129 relevance, 34, 35, 102, 159, 162, 165, 172, 214 rephrase, 118, 134, 146, 199 rhetoric, 59, 73, 79, 86, 87, 162 rhetorical situation, 87 The New Rhetoric, 88 S Sapir-Whorf hypothesis, 9 Scriven, Michael, 159 Searle, John, 83 self-defeating statements, 132 semantic incompatibility, 133, 148, 180, 186, 205 semantics, 41, 68, 79, 80, 87, 89, 92, 93, 119, 132, 178, 203 formal semantics, 178 informal argument semantics, 153, 163, 166, 178, 180, 181, 193, 219 semantic shifts, 92, 211 Snoeck Henkemans, A. Francisca, 35, 165 speech act theory, 76, 83, 164 Spengler, Oswald, 138

Index Sperber, Dan, 32, 34, 121, 159, 172, 192 standpoint, 57, 70, 83, 168, 196, 212 Stevenson, Charles, 130

T tautology, 127, 140 theory of mind, 7 thought, 3, 4, 6, 11, 47, 214 Tindale, Christopher, 48, 99, 102, 115–119, 122, 123, 125 Toulmin, Stephen, 52, 146, 157, 165, 177, 197

V vagueness, 73, 119, 134, 148, 158, 163, 180, 193, 198, 210 validity, 60, 80, 98 van Eemeren, Frans, 42, 57, 63, 67, 69, 71, 80, 82, 85, 98, 102, 105, 107, 117, 158, 160, 161 van Laar, Jan Albert, 117 Vatz, Richard, 87 verbal disputes, 89, 118

W Wagemans, Jean, 103, 170, 174, 176, 197, 202 Walton, Douglas, 35, 36, 52, 55, 61, 64, 67, 71, 85, 102–104, 110, 126, 174, 176, 219 warrants, 53, 157, 175, 209 Whately, Richard, 98, 101, 114 Wilson, Deirdre, 32, 34, 121, 159, 172, 192 Wittgenstein, Ludwig, 4, 13, 24, 28, 49, 139, 143, 144, 147, 190 Wohlrapp, Harald, 59, 69 Woods, John, 99, 126, 154

Z Zenker, Frank, 165, 177