Lying at the Semantics-Pragmatics Interface 9781614510840, 9781614510925

Table of contents :
Contents
Preface
1 Perspectives on lying
1.1 Introduction
1.2 Lying in the private sphere, on the Internet, and in politics
1.2.1 Lying in the private sphere
1.2.2 Lying on the Internet
1.2.3 Lying in politics
1.3 Ethical versus linguistic perspectives
1.4 The acquisition of lying
1.5 Psychological and neurological findings
1.6 Universality of lying
1.7 Lying and deceiving
2 Approaching a linguistic concept of lying
2.1 Introduction
2.2 The semantics-pragmatics distinction
2.3 Lexical semantics: The meaning of to lie
2.4 Interface approach: Lying and sentence types
2.5 The case of relative clauses
2.6 Conclusions
3 Assertion and lying
3.1 Introduction
3.2 Assertion and statement
3.3 The sincerity condition
3.4 Perlocution
3.5 The maxim of Quality
3.6 Assertoric commitment
3.7 Truth, truth conditions, and truthfulness
3.8 Assertion, common ground, and presupposition
3.9 Indirect assertions and explicit performatives
3.10 Expressive meaning
3.11 A simple definition of assertion
4 Defining lying
4.1 Introduction
4.2 A definition of lying
4.3 Intention to deceive
4.4 Bald-faced lies and self-deception
4.4.1 Bald-faced lies
4.4.2 Self-deception
4.5 Lying and falsely implicating
4.5.1 Normativity and cancellability
4.5.2 A definition of conversational implicature
4.5.3 “Lying while saying the truth”: Extended definition of lying
4.6 Falsely presupposing and falsely conventionally implicating
4.7 Underdeterminacy and explicature/impliciture
4.8 Vagueness and imprecision
4.9 Prosocial lies
4.10 Conclusions
5 Interface Problems
5.1 Introduction
5.2 Aspect-sensitivity, or: Did Clinton lie?
5.3 Fiction
5.4 Irony and tautology
5.5 Understatement and overstatement (hyperbole)
5.6 Lying and bullshitting
5.7 Conclusions
6 Lying and quotation
6.1 Introduction
6.2 Lying, direct/indirect quotation, and the verbatim assumption
6.3 Lying and mixed quotation
6.4 Lying and scare quotation
6.5 Fairness and transparency
6.6 The case of the German quotative wollen
6.7 Conclusions
7 Relative lies
7.1 Introduction
7.2 Accuracy and disagreement
7.3 Verum focus
7.4 The assertion of clarity
7.5 The assertion of certainty
7.6 Reliability and selfless assertion
7.7 Conclusions
Epilogue
References
Index

Citation preview

Jörg Meibauer Lying at the Semantics-Pragmatics Interface

Mouton Series in Pragmatics

Editor Istvan Kecskes Editorial Board Reinhard Blutner (Universiteit van Amsterdam) N.J. Enfield (Max-Planck-Institute for Psycholinguistics) Raymond W. Gibbs (University of California, Santa Cruz) Laurence R. Horn (Yale University) Boaz Keysar (University of Chicago) Ferenc Kiefer (Hungarian Academy of Sciences) Lluís Payrató (University of Barcelona) François Recanati (Institut Jean-Nicod) John Searle (University of California, Berkeley) Deirdre Wilson (University College London)

Volume 14

Jörg Meibauer

Lying at the Semantics-Pragmatics Interface

Funded by Volkswagen Foundation

ISBN 978-1-61451-092-5 e-ISBN 978-1-61451-084-0 ISSN 1864-6409 Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.dnb.de. © 2014 Walter de Gruyter, Inc., Boston/Berlin Typesetting: PTP-Berlin Protago-TEX-Production GmbH, Berlin Printing and binding: CPI books GmbH, Leck ♾ Printed on acid-free paper Printed in Germany www.degruyter.com

Contents Preface | ix 1 1.1 1.2 1.2.1 1.2.2 1.2.3 1.3 1.4 1.5 1.6 1.7

Perspectives on lying | 1 Introduction | 1 Lying in the private sphere, on the Internet, and in politics | 2 Lying in the private sphere | 3 Lying on the Internet | 10 Lying in politics | 12 Ethical versus linguistic perspectives | 14 The acquisition of lying | 15 Psychological and neurological findings | 17 Universality of lying | 22 Lying and deceiving | 25

2 2.1 2.2 2.3 2.4 2.5 2.6

Approaching a linguistic concept of lying | 33 Introduction | 33 The semantics-pragmatics distinction | 34 Lexical semantics: The meaning of to lie | 41 Interface approach: Lying and sentence types | 47 The case of relative clauses | 59 Conclusions | 62

3 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 3.11

Assertion and lying | 65 Introduction | 65 Assertion and statement | 65 The sincerity condition | 71 Perlocution | 74 The maxim of Quality | 76 Assertoric commitment | 79 Truth, truth conditions, and truthfulness | 83 Assertion, common ground, and presupposition | 88 Indirect assertions and explicit performatives | 91 Expressive meaning | 94 A simple definition of assertion | 99

vi   

   Contents

4 4.1 4.2 4.3 4.4 4.4.1 4.4.2 4.5 4.5.1 4.5.2 4.5.3 4.6 4.7 4.8 4.9 4.10

Defining lying | 102 Introduction | 102 A definition of lying | 102 Intention to deceive | 104 Bald-faced lies and self-deception | 107 Bald-faced lies | 107 Self-deception | 109 Lying and falsely implicating | 113 Normativity and cancellability | 114 A definition of conversational implicature | 120 “Lying while saying the truth”: Extended definition of lying | 123 Falsely presupposing and falsely conventionally implicating | 137 Underdeterminacy and explicature/impliciture | 140 Vagueness and imprecision | 148 Prosocial lies | 152 Conclusions | 154

5 5.1 5.2 5.3 5.4 5.5 5.6 5.7

Interface Problems | 156 Introduction | 156 Aspect-sensitivity, or: Did Clinton lie? | 156 Fiction | 158 Irony and tautology | 161 Understatement and overstatement (hyperbole) | 168 Lying and bullshitting | 171 Conclusions | 182

6 6.1 6.2 6.3 6.4 6.5 6.6 6.7

Lying and quotation | 183 Introduction | 183 Lying, direct/indirect quotation, and the verbatim assumption | 185 Lying and mixed quotation | 195 Lying and scare quotation | 199 Fairness and transparency | 203 The case of the German quotative wollen | 206 Conclusions | 210

7 7.1 7.2 7.3

Relative lies | 211 Introduction | 211 Accuracy and disagreement | 211 Verum focus | 215

Contents   

7.4 7.5 7.6 7.7

The assertion of clarity | 219 The assertion of certainty | 223 Reliability and selfless assertion | 225 Conclusions | 232

Epilogue | 234 References | 238 Index | 252

   vii

Preface A truth that’s told with bad intent beats all the lies you can invent. William Blake, Auguries of Innocence

This book is a linguistic study of lying. Lying is conceived of as a speech act in which the liar has the intention to deceive the addressee about the facts and about their own beliefs. This is, of course, an analysis that is shared by many researchers. What is new in my approach is that lying is viewed as an act that comprises deliberately false implicatures and presuppositions. This liberal approach to lying is elaborated in this book at length. It demands an engaged discussion of the relevant concepts, both from the perspective of linguistics and the philosophy of language. In particular, lying is located at the semantics-pragmatics interface, since it has to do with the manipulation of truth as well as with speech acts and pragmatic inferences. Thus, a number of theoretical approaches dealing with the semantics-pragmatics distinction are discussed with respect to lying. Because linguistics in general and pragmatics in particular are seen as cognitive abilities, psychological findings about lying are also taken into account. This book has been long in the making. Looking back, I gratefully remember a seminar on speech-act theory given by Marga Reis in 1976. This seminar certainly was the seed for my general interest in speech acts. My interest in lying started with an invitation to the conference “Cultures of Lying” that was held in Regensburg in 2002. I am grateful to Jochen Mecke for inviting me to this conference. Since then I have thought about the relation between asserting, lying, and falsely implicating. I managed to lecture about this topic at my home university, the Johannes Gutenberg University in Mainz in the winter term 2007/2008. However, because of other academic duties there was no time to work on a book manuscript and so I am very grateful for a grant from the Volkswagen Foundation that allowed me to focus on this “opus magnum”. I would also like to thank the anonymous reviewers and the editor of this series, Istvan Kecskes. Many thanks go to Kathrin Lakeberg who had a keen eye on my English. On a final note, I would like to thank the people I worked with most closely during the last few years, especially Hans Altmann, Franz d’Avis, Rita Finkbeiner, Bettina Kümmerling-Meibauer, and Markus Steinbach, and most of all my family, Erika, Gustav, and Bettina. November 2013

Jörg Meibauer

1 Perspectives on lying 1.1 Introduction Lying is a topic everyone is interested in. Being a liar and being lied to are fundamental experiences in human life. Lying can be approached from various perspectives: from ethics and religion to pedagogy and jurisprudence, from novels and films to theatre and photography. A great tradition of analyzing the phenomenon of lying exists in philosophy and the philosophy of language, with St. Augustine, Kant, Schopenhauer, and Nietzsche as the most famous thinkers. However, reflecting on the fact that lying is primarily a verbal act, one wonders why there is so little linguistic analysis of lying. In linguistics, we have a long tradition dealing with truth and truth conditions, but only a few attempts at clarifying the speech act of lying. Suppose I am lying on the beach. My friend Ken comes along and after greeting each other, I ask him why Barbie did not come with him. Ken replies, using air quotes, that for Barbie the sun is “too dangerous”. But that is not exactly what she said (as a witness told me). In reality, she said that she had to be careful because the burning sun was not good for her sensitive skin. What do you think, did Ken say the truth? Or was he manipulating the truth in a certain way, making you think that Barbie was hypersensitive? There is no easy answer to this question. We simply have the feeling that we should know much more about Ken, Barbie, and the context before we are able to evaluate Ken’s utterance. That we are unsure is, I think, part of a more fundamental problem. While we do have basic criteria of what lying or deception is, it is difficult to apply these criteria when it comes to utterances that are situated in a grey area such as in our introductory scenario. This book tries to shed some light on this grey area, using analytical tools developed in research on the semantics-pragmatics interface. Drawing the boundary between semantics and pragmatics – both being disciplines that deal with linguistic meaning – is one of the most basic problems of modern linguistics. In recent years, a lively debate has emerged from this problem (cf. Bianchi 2004, Szabó 2005, Allan and Jaszczolt 2012). Most researchers engaged in the debate relate their approaches to the fundamental work of Paul Grice, who made the by now classical distinction between ‘what is said’ and ‘what is implicated’ (Grice 1989a). By and large, ‘what is said’ is a semantic notion related to literal, context-independent meaning, while ‘what is implicated’ refers to meanings arising in the context of an utterance, and is therefore a pragmatic notion. Whether this distinction is necessary and whether it is sufficient for drawing the

2   

   Perspectives on lying

boundary between semantics and pragmatics is a matter of debate itself. We will have a closer look at this debate in section 2.2. Every one of us is occasionally being lied to (something we usually do not like) and, yes, everyone is a liar too. So while in general, we do not like lying very much, we accept it and find it very useful for practical purposes. How can we explain this astonishing flexibility in the acceptability of lying? Throughout this book, I will argue that this flexibility has not only to do with well-known moral issues (or hypocrisy, if you want), but has also to do with matters of how our language works. For instance, in the above example, Ken’s speech report plays an important role when trying to find out whether he is a liar, but we do not know much about the relation between lying and indirect speech. Hence this book pursues a double strategy: On the one hand it tries to analyze lying by situating it within the broader debate on the semantics-pragmatics interface; on the other hand it wants to show that this debate, usually dealing with a rather restricted set of data, may profit from a study of a “big” issue like lying. It is obvious that lying is indeed a case that is suitable for this purpose. It has to do with truth and truth conditions, i.e., issues that are traditionally associated with (truth-conditional) semantics. But lying is also a speech act that is deeply embedded in rich situational and discourse contexts. What a lie is cannot be found out without considering the cognitive and social goals the liar has in mind.

1.2 Lying in the private sphere, on the Internet, and in politics It goes without saying that we are most sensitive to lying in the private sphere, in our everyday face-to-face communication. It may happen between only two speakers such as between Ken and me in our initial example. But lying happens as well in other areas of communication, for example in computer-mediated communication or politics. What distinguishes these domains is the situational context that is accessible to us. While in face-to-face communication, speakers have direct access to a multitude of information about the participants’ intentions (e.g. facial expression, voice, proxemic and kinesic information), this information is typically lacking in computer-mediated communication and in reports on the lying behavior of politicians we come across in the media. In the following section, I will sketch some aspects of these three domains of communication by highlighting points that are of interest for our task of describing lying at the semantics-pragmatics interface.

Lying in the private sphere, on the Internet, and in politics   

   3

1.2.1 Lying in the private sphere Let us begin with lying in the private sphere. As a starting point, we take the following script from Saarni and Lewis (1993: 1–7). The authors propose “to think of it as similar to what television producers might work with as they try to imagine how effective it might be as a televised vignette.” (Have a look at the TV series “Mad Men”, for example.) Speech acts or thoughts that are lies or deceptions from the point of view of the authors are numbered consecutively and underlined.¹ It is clear that invented examples like these are not to be mixed up with authentic examples; however, as we will see in later chapters, there is a large tradition of arguing about lying while using stories that deliver the necessary contextual information, so this procedure is justified. Scene 1 SETTING: It is 6:30 AM in Jan’s and Ron’s bedroom. The radio has just turned on. No one turns off the alarm or stirs, but Ron’s blankets are in a heap while Jan’s are neatly drawn up to her chin. JAN: [Camera zooms onto her face.] Her eyes flick open toward the alarm clock and close again; her eyebrows knit together as in irritation and she pulls the blankets over her more tightly. A voice-over begins: VOICE-OVER: “He always thinks I should turn off the alarm, because if I complain about how hard it is for me to wake up in the morning, then all the more reason I should throw myself out of bed like some kind of automaton. [1] Well, I guess I’ll just have to sleep soundly….” RON: [Camera zooms onto his face.] A frown passes over his face. VOICE-OVER: “She really ought to push herself to get up and turn that damn thing off. As usual, I have to do everything around here.” [Camera backs off.] Ron staggers up, shits it off, and leaves for the bathroom. Now in front of his mirror, he examines his face, widening his eyes, and baring his teeth. VOICE-OVER: “Handsome devil, you. You’re not god’s gift to women, but you’ve been appreciated by quite a few [smiles an exaggerated lecherous smile], if I say so myself. I wish she did more of that appreciating [jerks head toward bedroom]. I could use some appreciating on another level from the ‘Honcho’ himself at work [grimaces]. He always wants more done, preferably all ready yesterday or even before he gives the order. As though we could all read his mind – if he has a mind. I wonder if Sharon is going to come today [smiles again]. Now I sure could do some appreciating of her!”

1 In the original text, these examples are starred. The numbers and underlining are my addition.

4   

   Perspectives on lying

JAN: [Camera shows her now sprawled across the middle of the bed, a slight smile on her face, eyes still closed.] VOICE-OVER: “How delicious these extra five minutes in bed are! These few minutes are about the only time I get to myself it seems, that’s why I need them – I don’t control anything around here! Not that he would understand; he thinks I spend the day watching television or curling my hair. No appreciation for what goes into taking care of two kids and being a freelancer, not to speak of taking care of him too. Let’s see, what’s my mental priority list for today… . RON: [Sticks head around side of bathroom door.] “Jan, get up! Is this one of those mornings, again, where you need coffee first, in bed, before you can deal with the world?” JAN: “As a matter of fact, I think it is one of those mornings. [2] Thanks in advance, honey” [said in a sweet tone of voice with just a hint of saccharin]. [Pulls blankets up tightly around her again.] “The girls will be in on top of me any minute now” [pulls blanket all the way over her face]. [Background noise of young girls’ voices and running feet. Camera switches to Ron in the kitchen pouring coffee with audible background evidence of the girls jumping on their mother and mutual happy morning greetings being exchanged.] RON: [With a resigned facial expression.] VOICE-OVER: Gotta put Sharon out of my mind. I love those girls, and I wouldn’t want to break up this family for the world.” Scene 2 SETTING: Jan is on her way to an appointment at a magazine publisher’s office; she uses the automatic change machine in the subway station in order to buy her ticket. Much to her obvious expressive pleasure, the machine dumps a huge handful of quarters, instead of the four that it should have, in exchange for her dollar bill. The money even spills out on the floor in front of the machine, making an attention-drawing clatter, as Jan scrambles to collect it all. Other passersby approach Jan. JAN: [3] “I just dropped my coin purse. What a way to start the day! Sorry for the noise” [smiles broadly at the two people closest to her]. Jan hurriedly leaves the area. Scene 3 SETTING: Ron is at work, and at the moment he is standing in the doorway to the office supply storage room with Sharon. They interact warmly: many smiles, head tilts, considerable eye contact, responsive body posture, etc. RON: “How old did you say your son was, Sharon?”

Lying in the private sphere, on the Internet, and in politics   

   5

SHARON: “He’s 7 now; wait, I have a snapshot of him here [digs in handbag and fishes it out]. He’s a real dear one, but awfully sensitive at times, you know, like he’ll come all upset over something some kid said to him. I guess he needs to develop a little more of a thick skin so he won’t be so vulnerable. On the other hand, god forbid he should be like his father! His ‘skin’ was so thick, it was like steel, and just as cold too, inside and out!” RON: “Must be hard being a single parent: Everything is your problem.” SHARON: “Yeah, but at least all I have is a kid to take care of and not a husband too.” RON: “Well, I take care of my wife, like bringing her coffee in bed in the mornings, and other things. [4] I’m a feminist, you know [said coyly].” SHARON: “Hmmm, I’ve never met a man who said he was a feminist unless he had some strategy in mind [said while looking sideways at Ron].” RON: “Oh, don’t misunderstand me! What I mean is that I believe in equal rights for men and women, equal pay, and all that sort of thing.” SHARON: “Who does the laundry in your family?” RON: [backing away a little bit from Sharon.] “My wife does. Look, I didn’t come to argue with you” [[5] irritation flickers across his face but is quickly replaced with a contrite look; then Ron brightens and changes the subject]. “What do you think the Honcho is going to perpetrate on us at the next meeting?” [Camera switches to viewing Ron and Sharon walking down hallway, backs to the camera, a distinct distance between them. Then the camera zooms in on Ron, alone and appearing disgruntled, looking out the window from his office.] RON VOICE-OVER: “Did I blow it or did I blow it? [6] But she’s not worth the hassle. In fact, I’ll bet she even kind of enjoyed needling me about that laundry thing. Toxic woman, that’s what. [7] Good thing I’m a family man and have my values clear.” Scene 4 SETTING: A board meeting at the advertising agency where Ron and Sharon both work; presiding is Mr. Lyecourt, otherwise known as the Honcho. The topic is an evaluation of how successful certain kinds of ad strategies are for selling a product. SHARON: [Alternately looking serious when turned toward her co-workers and smiling artfully when addressing the Honcho.] “Our television perfume ads were designed to subtly arouse, and while viewers were in the aroused state, to design [8] messages that would suggest they would be as sexually alluring as the model wearing the perfume. We think it worked: compared to their levels before the ads started in the targeted area, sales of ‘Compulsion’ and ‘Essence of Me’ more than doubled. Our market survey revealed that the profile of the typical buyer was

6   

   Perspectives on lying

female, worked outside the home, watched television most evenings, and considered shopping a pleasurable pastime.” [She sits back in her chair and directs a confident smile around the room.] MR. LYECOURT: [Initially looks approvingly at Sharon but then furrows his brow as he contemplates the remaining ad executives.] “So, why don’t the rest of you learn to do the same! You’ve got to figure out how to hit the viewers where they feel it: in their groin, in their pocketbook, their looks, their status; what they mean to other people in their lives. The worst thing a person can feel is that he’s trivial, meaningless, a bit of mold on a wall that can be washed off with a flick of a rag. Next to feeling this bad is just feeling ordinary. They always need to feel they’re somehow special, whether it’s their sex appeal, their brains, or their muscles. So go for it! Design ads that persuade people how to surpass ordinariness. That’s what sells. After all, it’s the American way, you know – the rugged individualist is ultimately a celebrity, because he’s special, and you know how everyone wants their claim to fame.” [He sits back pompously and with an assurance that he’s right.] Scene 5 SETTING: Jan sits hunched in front of her computer in a cramped, makeshift home office. She stares blankly ahead, and the screen on her monitor is also noticeably blank. JAN VOICE-OVER: “I know I’ve got the ideas; they’re hiding today or something. I have to keep reminding myself to be true to my own creativity and not get sucked into writing schlock, although it would be nice to make some money. There’s so much trash out there that people are just dishing out because it’s what sells or because it’s what you have to do to get some recognition. Murder, mayhem, wild sex, child abuse, and international drug cartels – that’s what’s in nowadays. Oh, and let’s not forget visions of Elvis and how to lose 50 pounds by going on an allchocolate diet.” [Telephone rings in the background; Jan sighs and lets the answering machine respond, but she jumps up with alacrity, coffee flying all over her keyboard, and runs to the phone when she hears who is speaking on the other end.] “Why, hello, Edna. [9] I was just in the middle of something rather demanding and couldn’t get to the phone right away … So you’re looking for something on extramarital sex and the lies that people tell about it … They would only pay $1500, hmmm … When would it be due? Well, that would be kind of rushing it. [10] Hang on, let me check my calendar here … Yes, I could do it, but I want an additional week … [11] Yeah, I have a major project I’m trying to finish up … [12] No, I’d rather not reveal what it’s about because I’m in the middle of some delicate negotiations around it …Glad you understand … Yeah, I know being an agent is tough stuff …Ok, send me the paperwork on it … .”

Lying in the private sphere, on the Internet, and in politics   

   7

[A couple of minutes later, Jan phones up Ron at work.] “Hey, guess what! I got a real juicy assignment … Ha-ha, you never guess what it’s on … This could be a real breakthrough for me, like I think this topic will really get noticed and get me some more contracts … [13] No, I don’t mind writing about this at all; hey, it’s what sells and gets a writer a little name recognition, sort of like your job, advertising.” Scene 6 SETTING: It’s evening: Ron is looking at television and Jan is paying bills. Their daughters, 7-year-old Phoebe and 9-year-old Kate, are sprawled on the floor in the family room playing a board game and also involved in some kind of dispute. KATE: [Looking disdainfully at her younger sister] “You are such a big baby, Phoebe, you always cry when you lose. You’re no fun to play with!” PHOEBE: [Looking distressed] “But you always beat me; I think you’re a cheater!” KATE: “No, I’m not; you are just dumb.” RON: “Kate, that really is out of bounds. I want you to apologize to Phoebe RIGHT NOW! And, Phoebe, you can’t take these games so seriously; [14] losers don’t cry, they just grin, and bear it and think about how to outfox the other player next time. Ok, Kate, let’s hear the apology.” KATE: [Averting her face from her father’s view and looking quite contemptuously at her sister.] [15] “Sorry.” JAN: [Noticing Kate’s facial expression toward her sister and her decidedly unenthusiastic tone of voice.] “Kate, do it again and this time mean it!” KATE: [Adopting the briefest of thin-lipped smiles and a somewhat high-pitched tone of voice] “Sorry.” RON: “That’s better. Why don’t you guys get out your paints and I’ll join you as soon as this program is over. By the way, Jan, that Sharon really alienated everyone today at our department meeting with old Lyecourt. Boy, she can be really underhanded.” JAN: “I thought you liked her and thought she was a really good addition to the group.” RON: “Guess it was just a first impression: Manipulation comes in all kinds of packages. [16] Now, if I had a team-player like you, I’d by flying high at work!” JAN: “Why, Ron, that’s really sweet of you to say that!” [The scene fades out and the television producer wonders what to do with the script. She files it away under the old soap opera series, “As the World Turns,“ and mutters to herself, “It’s all so predictable.”] As said before, though this is an invented script, it is well-suited for pointing out some aspects of an everyday concept of lying. So let us see how the underlined examples might be classified. In their simple “taxonomy of lying”, Saari and Lewis (1993: 8–13) draw a distinction between three types of deceptions: “(1)

8   

   Perspectives on lying

ordinary deception toward others committed with self-awareness; (2) deception toward others that requires some degree of self-deception; and (3) self-deception even in the absence of another, that is, the need for illusion.” Table 1: A simple taxonomy of deception (Saari and Lewis 1993)

deception

ordinary deception

ordinary deception plus self-deception

self-deception

While it may be apt to include self-deception into a taxonomy of deception, it is evident that self-deception is quite an intricate concept. This becomes clear when we look at the following definition taken from Deweese-Boyd (2006:  1): “[…] self-deception is the acquisition and maintenance of a belief (or, at least, the avowal of that belief) in the face of strong evidence to the contrary motivated by desires or emotions favouring the acquisition and retention of that belief.” Because our focus is the communicative act of lying, we refrain from considering self-deception here: Whatever the exact definition is, self-deception may happen without any communication. (We will come back to the notion of self-deception in section 4.4.2.) Furthermore, while it is intuitively plausible that lying includes deception, it is not clear that deception includes lying. For example, it is possible to deceive without any utterance, just by acting in a certain way. Hence we need to distinguish between lying and deceiving, a task Saari and Lewis appear not to be interested in. In example [1], Jan pretends to be sleeping while really she is awake. Thus she deceives Ron with respect to her state. But [1] is only a thought that accompanies her acting as if she is asleep, and therefore does not count as a lie. Lying is usually conceived of as a verbal act. Example [2] could be ironical – that is, Jan is not really grateful because she knows that Ron does not like that job. Irony, however, has to be distinguished from lying. Example [3] displays a genuine lie. In this situation, it is not true that Jan just dropped her coin purse, and Jan knows this. Example [4] is a good candidate for the second type of Saari and Lewis’ simple classification, namely “deception toward others that requires some degree of selfdeception”. Here, Ron asserts that he is a feminist, and this is (at least according to the lexical definition of feminist) not true. On the other hand, if bringing your

Lying in the private sphere, on the Internet, and in politics   

   9

wife coffee to bed is supposed to be indicator of “feminism”, there may well be a sort of self-deception involved. This example shows that it is sometimes not clear whether a proposition is true or false. In example [5], Ron shows contrition but is not in the corresponding psychological state. Thus, he deceives Sharon about his true feelings. Nevertheless we have a case that does not qualify as a lie because the deception is on the level of the mimics. In the voice-over examples [6] and [7], we see that Ron is adapting to the new situation. The respective thoughts may be untrue (Ron was toying with an affair, and thus he is not a “family man”), but they could also qualify as instances of self-persuasion. Example [8] refers to the fact that deceptive acts are typical for advertisements. Here, it is wanted that the customers associate a certain positive image with the product that they are supposed to buy. Jan’s utterances in [9]–[12] serve the overall goal to portray herself as a very busy and demanded author. Examples [9], [11], and [12] are plain lies. [10] is true on the literal level, however, it is intended to implicate that Jan has a full calendar which is – as far as her job as a freelancer is concerned – not true. I will later argue that deliberately false implicatures should be considered as lies, too (see section 4.5). When Jan asserts in [13] that she likes to write on the topic of extramarital sex, this may be an outright lie or a case of self-deception. Whether [14] is a lie or a deceptive act, is quite hard to tell. It is not so much a statement of facts (“It is true that all losers behave in such and such way…”), but rather a stereotypical idea on the behavior of losers. It comes close to bullshit, i.e., misleading utterances in which truth does not really matter to the speaker. (We will come to the discussion of bullshit in section 5.6.) Example [15] shows insincerity, which means showing emotions that are not actually felt. Lies are insincere assertions, as we will see later, whereas insincerity is a sort of mental reservation that is not specific for lies – as, for example, promises may be insincerely uttered, too. In the final example [16], Ron pays Jan a compliment but it is not clear whether he really means it, or whether he is being deceptive. Remember his fantasies about Sharon in the morning. These scenes nicely show the emotional quality of the thoughts and utterances, and the ubiquity and pervasiveness of deceptive acts in everyday speech. Furthermore, they show that it is a difficult task to properly distinguish between lies and deceptions, and between other-deception and self-deception. At this point we need some fine-grained definitions.

10   

   Perspectives on lying

Note that we did not find any examples of the third type in Saari and Lewis’ taxonomy, namely illusion. Illusion, understood as latent self-deception, is certainly not a verbal act and therefore not in the focus of our interest. What is interesting to us, however, is bullshit. We already assumed that [14] might be an example of bullshit. In addition, recall the speech of Mr. Lyecourt in Scene 4. Does Mr. Lyecourt really mean what he says? One could argue that he is bullshitting about the American way and the wish to be something special. Bullshit is something that is close to lying but distinct from it.

1.2.2 Lying on the Internet Let us turn to the second domain of interaction, namely computer-mediated communication (CMC). In CMC, writers communicate via e-mail, social networking pages like Facebook, discussion groups, or blogs. Following Whitty and Joinson (2009), CMC is characterised by a truth-lies paradox. On the one hand people engaged in CMC show a tendency towards self-disclosure. They tend to tell more true details about themselves in comparison to face-to-face interaction. On the other hand people use the Internet extensively to lie about themselves, especially when CMC happens anonymously. The often-posed question whether lying happens more often on the Internet than in face-to-face communication is however a problematic one. Firstly, even in everyday face-to-face communication it is hard to tell how often people lie because this depends on an accepted definition of what should count as a lie. Secondly, it cannot be disregarded that lies in face-to-face communication are not evaluated on the basis of the same criteria as lying in CMC. As pointed out by Whitty and Joinson (2009: 56–60), lying in CMC is often connected with identity-based deception. People lie about age, gender, occupation, education, and income. In contrast, lying about age and gender is not that easy in face-to-face communication because of the visual information the participants have. In one study by Whitty (2002) based on 320 chat rooms, she found that “men were significantly more likely than women to lie about gender, occupation, education and income” and “younger people (aged 17–20 years) were more likely than older people (aged 21–55 years) to lie about their age”. Identity-based deception is also typical of scams (like the famous ‘Nigerian e-mail’ scam, the lottery scam, and the romance scam), phishing, and spam. As for content-based deception in CMC, Whitty and Joinson (2009: 65–68) discuss whether there is more lying in CMC in comparison to other electronic media, and whether there is a difference with respect to self-serving versus other-serving lies. While so-called ‘social distance theory’ predicts more lies in

Lying in the private sphere, on the Internet, and in politics   

   11

rich media than in face-to-face communication, because less lying makes the liar feel less uncomfortable, the so-called ‘media richness theory’ would argue that richness of informational cues elicits lying. As face-to-face communication is very rich of information, it is precisely here that a large amount of lying is to be expected. According to the first approach, a possible lying-likeliness-scale could be e-mail > Instant Messenger > phone > face-to-face communication; according to the second approach, the scale would be face-to-face communication > phone > Instant Messenger > e-mail. However, research by Hancock et  al. (2004) reported in Whitty and Jones (2009) appears to have found that lying happened more often during telephone calls than in e-mail. Hancock et al. (2004) conducted a diary study with 28 students reporting on their lying behavior for seven days. In face-to-face communication, 202 lies were recorded, followed by the telephone (66 lies), Instant Messengers (27 lies), and e-mails (9 lies). However, it is unclear to me whether these data have been matched with the input of the students. For instance, one particular student had only 60  minutes of face-to-face communication per day while he was talking on the telephone for four hours a day; on another day it was the other way around. It is very likely that students differ in their communicative behaviour, just as they differ in their individual output. So there is much more to be said about pragmatic productivity and occurrences of lying. Another issue is the criteria the students apply when it comes to judging what a lie is. From experimental studies on the notion of lying we know that there is quite some variation in this area (see section 2.3). Hancock et al. (2004) go on and offer a further approach, called the ‘feature based theory’. This approach considers whether a medium is synchronous, unrecorded, and whether participants are co-present. The basic idea is that people tend to lie more often the more synchronous a medium is, because lying typically happens spontaneously. Moreover, liars prefer unrecorded media, because they fear later detection if the communication is recorded. Finally, in “distributed” communication, when participants are not co-present, it is easier to lie, for instance, with respect to the actual behavior of the liar (e.g. asserting to work while in fact playing). Thus, liars are better able to calculate their risk of being caught in the act of lying. As Whitty and Joinson (2009) report, there appears to be a different lying behavior with respect to self-serving versus other-serving lies. Self-serving lies are avoided in face-to-face communication because liars feel uncomfortable when lying to participants that are close to them, while they are more easily made in CMC. Other-serving lies that protect the face of someone close to us are probably more apt to face-to-face communication because it is easy to tell the truth with respect to strangers in CMC. If these hypotheses are true, Whitty and Joinson

12   

   Perspectives on lying

(2009: 67) argue, this can be seen as evidence for the social distance theory and against the feature based theory. Participants in a study done by Whitty and Carville (2008) said that other-serving lies to strangers are preferred in face-to-face communication rather than in e-mails. The feature based theory, as we have seen, argues that not being co-present facilitates lying. Again, this shows that the distinction between these types of lying is an important one. Cheating on online dating sites and Internet infidelity are further deceptive activities that play a role in everyday life (Whitty and Joinson 2009: 73–96) and thus deserve further empirical exploration. I think it is important to keep ongoing lying activities in numerous types of CMC in mind when dealing with the topic of lying at the semantics-pragmatics interface.

1.2.3 Lying in politics Finally, consider lying in the political sphere. From the perspective of the average citizen, lying in the political sphere is mediated by the media (newspapers, radio, TV, Internet, etc.). In most cases, the individual member of society has no direct access to the lies of a politician, and even if a lie is directly accessible (i.e., if a lie occurs as part of a politician’s speech) it is hard to interpret it because of a lack of relevant contextual information. For instance, the French Prime Minister Nicholas Sarkozy falsely asserted in September 2010 that Germany planned to deport Roma into special camps. The German chancellor Angela Merkel denied this but the motives for his wrong assertion were not uncovered. What would count as an outright lie in the private sphere is often downplayed as a case of misinformation in the political sphere. This has to do with the strong moral reservations against lying in a liberal democracy (cf. Cliffe, Ramsay, and Bartlett 2000). Since 2000, there are many lying scandals to be reported, each worth a closer examination. Here are just three of them: – “[…] in an interview given to the American ABC television network on February 9, 2005, former US Secretary of State Colin Powell admitted that his United Nations Speech making the case for the US-led war on Iraq was a painful blot on his record. On February 5, 2003 he had addressed a plenary session of the United Nations Security Council to argue in favour of military action. Citing ‘numerous’ anonymous Iraqi defectors, Powell asserted that there could be no doubt about Saddam Hussein possessing biological weapons and his capability of rapidly producing more” (Mecke 2007: 1). The Iraq War 2003 and the roles of George W. Bush and Dick Cheney are discussed in detail in Carson (2010: 212–221, 224–227). Carson (2010: 227) concludes: “Lying and deception

Lying in the private sphere, on the Internet, and in politics   

–

–

   13

by the Bush Administration brought about a rush to war when war should have been the very last resort.” In 2006, in the face of elections, the Hungarian Prime Minister Ferenc Gyurcsany lied about the real situation of the national economy. Later, at a meeting of the socialist party at Balatonöszöd, he confessed that he had lied: “Naturally, we lied throughout the last 18 months, two years. It was totally clear that what we were saying was not true. […] We lied in the morning, we lied in the evening.” The publication of this speech, being secretly recorded and sent to a broadcasting station, led to demonstrations against him (Mecke 2007: 1). In 2011, the German Minister of Defence, Karl Theodor Freiherr zu Guttenberg, stepped down because his Ph.D. dissertation on European constitutional law turned out to be a case of plagiarism. He was accused as a liar and swindler, there were protest letters by German academics, and even a website engaging in the precise documentation of the amount of plagiarism was created. Guttenberg admitted that his dissertation was full of errors, but he always insisted on not having intentionally betrayed anyone. This statement, however, was regarded as an outright lie by his critics.

As in the private sphere, the motives for lying are various and often complex. Carson (2010: 208) points out that most cases of lying and deception in the political sphere fall into “one or both of the following categories: 1. Lying and deception to manipulate public opinion and generate support for actions, causes, policies, and political objectives that one wants to promote. 2. Lying and deception to promote one’s personal interests, e.g., a politician lying to avoid impeachment or increase his chances of winning an election.” An example for the first category is the Iraq War (Bush Administration), examples for the second category are the cases of Gyurcsany and Guttenberg. While there is a certain tendency in ethics towards a more descriptive view of lying, Carson (2010: 257–265) holds the view that honesty is still a virtue: “Lying and deception by political leaders are great betrayals of trust in democratic societies: they thwart or subvert the will of the populace. There is a very strong presumption against doing this in a democratic society (…).” Although I would like to see more empirical research concerning the public reflection of these attitudes in CMC discussion groups, the very fact that there are such discussions points into the direction that traditional notions of honesty exist and are alive. Having pointed out aspects of lying related to the private sphere, the Internet, and politics, we now turn to ethical perspectives on lying, as opposed to the linguistic aspects that are the essential topic of this book.

14   

   Perspectives on lying

1.3 Ethical versus linguistic perspectives It goes without saying that ethics and linguistics are separate, autonomous disciplines. Yet there is a long-standing tradition of viewing the speech act of lying as something that has to do with morality. Generally, lying is considered to be a behavior that should be condemned because it is bad. In recent ethical reasoning, lying is still on the agenda. Broadly speaking, there are two approaches to the matter. The first approach, elaborated, for instance, by Dietz (2002) portrays lying as a sort of rational behavior, as it serves certain social goals of the liar. And while there are certainly liar’s goals that would be rated as damnable by most people, there are others that, on the contrary, are praised. For instance, if someone lies to a dying mother about the recent death of her only child, this would be rated as an acceptable if not necessary lie by many speakers. According to Kant (1968), however, lying can never be accepted because truthfulness is a human duty. The second approach points out the rational side of lying, while also tracing back the common aversion to lying to certain virtues. For Carson (2010), this is the cardinal virtue of honesty, or more precisely, “honesty in a negative sense”, i.e., “having a strong principled disinclination to tell lies or to deceive others” (Carson 2010: 257). (Honesty in a positive sense, by contrast, includes not only a disinclination to tell lies or to deceive others but also a candid and open attitude as well as the willingness to reveal information.) Since our concern is the linguistics of lying, in particular the relationship between semantic and pragmatic aspects of lying, ethical aspects of lying are not in our focus. However, it appears that they cannot be dismissed completely because there are at least two points of contact between linguistics and ethics. The first is the definition of illocutionary acts via felicity conditions which is the usual procedure in speech-act theory. In many approaches, there is something like the Searlean “sincerity condition”, requiring a certain mental state of the speaker. Thus, in the case of assertions it is required that the speaker be sincere (see section  3.3). Other philosophers assume that assertions are connected to the responsibility of the speaker (Alston 2000) or that the speaker be obliged “to warrant the truth”, as Carson (2010) puts it. Assuming that such qualifications are necessary at all, we face the problem that these notions are ethically loaded. To act insincere, irresponsible, or to not warrant the truth when being obliged to are certainly behaviors connected with moral evaluation. The second point of contact is politeness. I take it that politeness is a pragmatic and sociolinguistic notion. When one sticks to the classical approach of Brown and Levinson (1987: 61) who define ‘face’ as “that public self-image that every member wants to claim for himself”, then lying may be conceived of as an

The acquisition of lying   

   15

attack against that face. In particular, lying is an attack against the hearer’s negative face, i.e., “his basic want to maintain claims of territory and self-determination” (Brown and Levinson 1987: 70). To disrespect face needs is also connected with moral evaluation. In the course of this book, I will not be able to pinpoint the exact nature of these points of contact between ethics and linguistics. What would be needed to do this is a very broad theory about the way in which our everyday behavior is guided by moral evaluations and virtues. Since such a theory is lacking, I will appeal to the Gricean principle of cooperation, a rational principle that demands cooperation in communication (cf. Grice 1989a). While lying and deceiving others is uncooperative in most cases, there are other cases where lying may be cooperative (see section 4.9). Similarly, while being impolite is certainly uncooperative in most cases, situations occur in which being impolite is an adequate, if not necessary behavior.

1.4 The acquisition of lying Since lying is a common human behavior, children have to acquire the skill of lying. (And if they do not, they are probably pragmatically impaired.) The ability to lie goes together with the acquisition of metalinguistic abilities; for instance, lies must be differentiated from ironies, metaphors, and jokes (cf. Leekam 1991), and this presupposes an ability to reflect upon language and its functions. A milestone for a child’s acquisition of lying is the ‘Theory of Mind’, understood as the ability to consider the thoughts, imagination, and feelings of others and respect them when acting. Experimental research has shown that these abilities are usually acquired by the age of four. Precursors of the Theory of Mind are the imitation of intended actions at the age of 1;6 years, the distinction between one’s own and others’ feelings or goals, and the onset of symbolic and fictional play. At the age of 2;0, the ability to ascribe feelings and wishes to others (independently of one’s own feelings and wishes) develops. Usually, a distinction between first-order beliefs and second-order beliefs is made. First-order belief is related to the understanding that one can have a false belief about reality (appearance-reality distinction) (3;6–4;0 of age), while second-order belief is connected with the understanding that one can have a false belief about the belief of another person (6;0 of age). The insight that different perspectives about a belief of another person are possible is a late achievement in the development of children and young adults (between 12;0 and 17;0 of age). A seminal methodological tool in the research of the Theory of Mind is the Sally-Anne-Task. Sally hides a toy in a basket (x) and leaves the room. Anne enters

16   

   Perspectives on lying

the room, takes the toy out of the basket, and puts it into a box (y). The question is where Sally will look when re-entering the room: in the basket (x) or in the box (y)? Children aged 2;6 to 3;0 answer that Sally will look in the box (y). From 3;6 to 4;0, children answer (correctly) that Sally will look in the basket (x) because this is where she left the toy. The assumption with regard to the acquisition of lying is that children initially do not grasp the aspect of intentionality; instead, they consider only factual truth or falsity. In a landmark study, Wimmer and Perner (1983) used the Sally-Anne-Task for research on the acquisition of lying. Participants were asked to indicate where the protagonist would look upon their re-entering. The results were as follows: “None of the 3–4-year old, 57% of the 4–6-year old and 86% of 6–9-year old children pointed correctly to location x in both sketches.” (Wimmer and Perner 1983: 103) A further question, aiming at the stability of children’s belief, was how children behaved when intending to deceive another person about the location of the toy, or with the intention to tell a friend the truth about the location of the toy. Here the result was: “Independent of age, of those children, who correctly thought that the protagonist would search in x, 85% of the time they also correctly thought that he would direct his antagonist to location y and his friend to location x.” (Wimmer and Perner 1983: 104) This shows that children are able to relate their representations of another person’s belief to successful acts of deception. Another influential research method is the Temptation Resistance Paradigm. Children are told that they should not spy or play with a toy. If later asked whether they spied or did play with the toy, they could either tell the truth or lie. Already 36% of the 3-year-old tend to lie while this behavior is evaluated as very bad by most of the 4-year-old (cf. Talwar and Lee 2008). Later in their moral development, children will detect that prosocial lies are different. When lying for reasons of politeness or to protect others, lying is rated as acceptable or even as a welcome social behavior. A further aspect directly related to children’s actual lying behavior is executive functioning, i.e., controlling one’s own lying history within a discourse. Generally, it is difficult for the liar not to mix up conflicting alternatives, e.g. what they have really done/thought vis-à-vis what they said they have done/thought. Higher mental and verbal abilities are demanded, for instance self regulation, planning of the discourse, flexibility of attention, discourse strategies, etc. These are abilities that develop only slowly. In the course of lying acquisition, the distinction between lying, mistakes, over- and understatements, and ironies is hard to grasp for most children. Until 6  years of age, many children mistake irony for lying, and even 9–13-year-olds have difficulties to understand the concept of irony (cf. Winner 1988, Creusere 2007, Bosco and Bucciarelli 2008).

Psychological and neurological findings   

   17

For more than 100 years, there has been research on the acquisition of lying (cf. Lee 2013, Lee and Talwar 2014). This research can tell us how the concept of lying arises in the mind of children. While the acquisition of lying is not in the focus of this book, I want to point out the relevance of lying acquisition for a general theory of lying with regard to two areas. The first area is intercultural pragmatics where it may be asked whether certain concepts of lying are more or less close to children’s conceptions of lying. The second area is experimental pragmatics where a lot of psycholinguistic results concerning the acquisition of scalar implicatures are available (cf. Katsos 2012). Here, one slogan is “children are more logical than adults” (cf. Noveck 2001), and there might be a parallel assumption that children stick more to factual truth than adults. Thus, developmental evidence may give insights into the lying concept of adults.

1.5 Psychological and neurological findings Lying is a mental activity, an achievement of the human brain. Hence lying is the subject of numerous experimental studies, often driven by the wish to get robust indicators of lying in legal cases. In this section, I will shortly touch upon several of the pertinent questions in this area of research, while trying to relate some of them to the topic of this book. It goes without saying that broad coverage of the literature is not possible here (but see Vrij 2008). One basic idea informing this research is that lying is cognitively more demanding than saying the truth. Under the pressure of being detected, the liar has to solve several tasks, and thus has to make more cognitive effort. Vrij (2008: 39–41) lists several factors that contribute to an increased mental load of the liar. Most importantly, the formulating of the lie may be demanding: “Liars may need to make up their stories and must monitor their fabrications so that they are plausible and adhere to everything the observer knows or might find out. In addition, liars must remember their earlier statements, so that they appear consistent when retelling their story, and know what they told to whom. Liars should also avoid making slips of the tongue and should refrain from providing new leads.”

Moreover, liars cannot take their credibility for granted, so the attempt to avoid everything that could further reduce their credibility adds to the mental load. Because of this, they possibly tend to monitor the addressee’s reactions more closely than the truth-teller. And last but not least, the concentration on their own acting and role-playing may add to cognitive effort, as well as the suppressing of the truth, and the strengthening of the intention to lie, since the truth is often activated automatically while sticking to the lie needs a special activation of the mind.

18   

   Perspectives on lying

Because of these factors that might influence the overall performance of the liar, one might look for clear indicators of lying, something akin to the growing nose of Pinocchio. Indeed, this is the hope of researchers that work with the polygraph. Among the nonverbal cues for lying ranges that have been researched in the past are the following vocal cues which I take from the synopsis of more than hundred studies reviewed by Vrij (2008: 90–99): hesitations, speech errors, highpitched voice, speech rate, latency period, pause durations, pause frequency. In addition, visual cues have been researched, such as gaze, smile, self-adaptors, illustrators, positions or movements of hand/finger, leg/foot, trunk, and head, as well as shifting position, and eye blink. The evidence is far more complex than can be shown here. Vrij (2008: 86–87) summarizes the state of the art: “People’s expectations that liars will show nervous behaviours, particularly gaze aversion and fidgeting, were not supported. Rather, liars gave the impression that they had to think hard or that they tried to control themselves. This results in cues such as word and phrase repetitions, long pauses, or a decrease in hand/finger movements.”

A further important result is that there is “no nonverbal behaviour that is uniquely associated with deception” (Vrij 2008: 88), leading the author to the assumption “that examining a cluster of nonverbal cues gives more valuable information to deception than examining each cue individually” (Vrij 2008: 89). Besides the nonverbal and the visual cues, there may be indicators in the verbal phrasing. It goes without saying that these devices are the most interesting from a linguistic point of view. I will come back to the question whether there are linguistic indicators for lying in chapter 2. Here, I will shortly report what Vrij (2008: 102–114) found. As candidates for promising verbal clues he lists the following (Vrij 2008: 102), which show up in a corpus of 69 studies (see the synopsis pp. 112–114): 1. Negative statements: statements indicating aversion towards an object, person or opinion, such as denials, disparaging statements and statements indicating a negative mood. 2. Generalising terms: the use of word such as ‘always’, ‘never’, ‘nobody’ or ‘everybody’. 3. Self references: the use of words referring to the speaker himself/herself such as ‘I’, ‘me’ or ‘mine’. 4. Immediacy: speaker’s response sounds direct, relevant, and clear rather than indirect, distancing, evasive. 5. Response length: length of the answer or number of spoken words. 6. Plausible answers: statements that make sense and that sound credible and reasonable.

Psychological and neurological findings   

   19

7.

Lexical diversity: number of different words in a statement divided by the total number of words used in that statement. 8. Consistencies: number of the same details repeated over two statements by the same speaker. 9. Contradictions: contradictions within a statement or between two statements. This list is by no means exhaustive. DePaulo, Lindsay et al. (2003) list many more. Again, the results of studies dealing with these putative cues are that an individual verbal diagnostic cue does not exist (Vrij 2008: 106). This is not surprising from a linguistic point of view; be it otherwise, the liar would commit “illocutionary suicide”, i.e., act in a paradoxical way. However, there seems to be a trend towards some of these cues, namely towards negative statements, generalizing items, self-references, and plausibility: In comparison to truth tellers, liars’ statements contained more of these elements although negative comments and plausibility had only weak effects according to the meta-study done by DePaulo, Lindsay et al. (2003). There are many psychological methods that aim at catching liars. Vrij (2008), in his excellent overview, mentions among these methods the Behaviour Analysis Interview, the Statement Validity Assessment, Reality Monitoring, Scientific Content Analysis, and the Polygraph, as used in the Concern Approach and the Orienting Reflex Approach. Moreover, he deals at length with functional Magnet Resonance Imaging (fMRI). Here I will shortly go into the measuring of event related potentials (ERP) and in fMRI, because these two methods are typically used in recent psycho- and neurolinguistic studies within the emerging field of experimental pragmatics. The two basic questions are: First, is lying, as measured in reaction times, more costly than telling the truth? This is exactly what one would think with regard to the fundamental assumption that lying is more difficult than telling the truth. Second, can the act of lying be located somewhere in the brain while doing it? If so, one would have an insight into the most basic cognitive abilities necessary for successful lying. The most basic approach of analyzing lying in the laboratory (and it goes without saying that this is an artificial communication situation) is to give the participants a Guilty Knowledge Task (GKT, also Guilty Knowledge Test, Concealed Information Test). The idea is that liars react with an orienting reflex when they become aware of a stimulus that is personally relevant for them. This orienting reflex is then measured by using either a polygraph or an EEG. A possibly examination scenario is given by Vrij (2008: 344):

20   

   Perspectives on lying

“Suppose that someone was killed with a knife which was left at the crime scene. As long as information about the crime has not been leaked out, only the murderer could know which type of knife was found at the crime scene. In a typical GKT examination, examinees are attached to the polygraph and will be shown several knives, including that used in the murder. For each knife the examinee will be asked, whether he or she recognises it as the one used in the crime. The examinee is instructed to answer “No” each time this question is asked. The murderer is familiar with the murder weapon and is therefore likely to show orienting reflexes when the knife used in the crime is shown; innocent examinees have no knowledge about which knife was used and will therefore show no orienting reflexes.”

For instance, in two reaction time experiments by Walczyk et al. (2003) participants were asked to lie or say the truth. The task was to answer personal questions either in a truthful or in a deceptive way. The study showed that the construction of a lie was time-consuming. Lying with respect to open or delicate questions was even more time-consuming. This is in line with the general assumption that lying requires more cognitive effort than truth-telling (see also Williams et  al. 2013). Note that Vrij (2008: 354–355) is very sceptical about the validity of reaction times “as a measurement of deceit in real-life settings”. It is problematic to rely on a single measure, the data being sensitive to countermeasures (e.g., when equalizing the answer pattern), and it is difficult to select the appropriate critical and filler alternatives. In ERP-studies, the typical approach is to measure electrophysiological signals related to the computation of texts. These wave patterns can be compared to wave patterns received in other studies, so that ultimately a fine-grained picture of the overall architecture of a text or an utterance may arise. Wave patterns associated with lying have been found in a number of studies. In particular, the so-called P300 (positive waves 300 ms after presentation of a cue) was identified (Vrij 2008: 357–362). P300-effects are to be seen when participants show an orienting reflex towards personally relevant stimuli. Hence, it is plausible that these effects occur regularly in Guilty Knowledge Tests. Another wave pattern was discussed in the recent study of Carrión, Keenan, and Sebanz (2010). The N450, a medial frontal negative deflection, is “a component sensitive to cognitive control demands and conflict processing that is thought to reflect activity in anterior cingulated cortex” (Carrión, Keenan, and Sebanz 2010: 106). The authors suspected that the medial frontal negativity (MFN) was sensitive to deceptive intent. They expected “this component to be more pronounced when people told the truth in an attempt to deceive others (choice condition) compared to when people stated the truth without deceptive intent (instructed condition)” (Carrión, Keenan, and Sebanz 2010: 106). However, the result was “that telling the truth with a deceptive intention created a similar demand on cognitive control processes as lying” (Carrión, Keenan, and Sebanz

Psychological and neurological findings   

   21

2010: 106): both were associated with a N450. They conclude that the intention to deceive is the important thing, not the truth or falsity of a statement. This conclusion supports some of the findings we will see in chapter 4 of this book. In their review of the neuroimaging of deception, Sip et al. (2008: 49) point out that interpersonal deception is a complex behavior that has to do with (a) information management (“keeping track of beliefs and monitoring feedback to maintain deception”), (b) impression management (“building trust through pseudo-cooperation and control of non-verbal signals”), (c) risk management (estimating the gains and losses of the deception and its long-term consequences”) and (d) reputation management (“justifying deception on the basis of the ‘greater good’”). The according operations are not deceptive in their nature but appear in processes of deception, and they are intertwined. In the fMRI-studies reviewed by Sip et  al. (2008), deception is associated with the dorsolateral prefrontal cortex (DLPFC), the medial prefrontal cortex (MPFC), and the adjacent anterior cingulated cortex (ACC). The DLPFC has to do with response inhibition and control, and this plays an important role when the liar monitors their responses. The MPFC is involved because mentalizing, understood as “the ability to read and manipulate the mental states of others” (Sip et al. 2008: 50), is also necessary for deceiving others. Finally, ACC is associated with risk management, decision making, and reward processing; in addition, it has also to do with “relating actions to their consequences, representing the expected value of actions and likelihood of errors”, but also “with problem solving, motivation and the anticipation of the consequences of action” (Sip et al. 2008: 50). While fMRI-studies certainly give evidence on these brain areas to be active during deception and lying, there is no such thing as a certain place or region in the brain solely responsible for deceptive behavior: “The problem is that deception involves many cognitive processes, none of which are unique to it” (Sip et al. 2008: 50). However, Harada et al. (2009: 32) recently suggest that the left temporo-parietal junction (TPJ) “might play a key role in the processing of both the conventionality and intentionality components of lie judgments”. Under the heading of “future improvements”, Sip et al. (2008: 50) propose to incorporate findings from other disciplines, explicitly naming pragmatics as a discipline dealing with intentional acts and contextual variation. Before we will come to a sketch of the semantics-pragmatics interface and the notion of context, we will shortly touch upon the question of the universality of lying.

22   

   Perspectives on lying

1.6 Universality of lying Lying is a kind of human behavior. We can safely assume that it is a property of the human species on a par with other cognitive abilities. Since lying appears to be connected with higher centres of the brain such as PFC (prefrontal cortex), we may assume that the ability to lie emerged during evolution. Why such an evolutionary output should occur is of course hard to tell. Adenzato and Ardito (1999) speculate that the evolution of deception has to do with group living. On the one hand, the group imposes restrictions on the individual member; on the other hand, the individual strives after some level of autonomy, and deceptive strategies may create an autonomic space. The same line of reasoning applies to the relation between a child and his or her parents. Lying could be seen as an act aiming at getting more control over one’s life and emancipating oneself from the restrictions of parents and the family. While lying may be fundamental from a cognitive and evolutionary point of view, it seems to be bound to social practices and, moreover, to diverse cultures. Hence we can ask whether there is a universal way of lying, a lying behavior that is uniform across cultures, or whether lying is different with respect to social groups, societies, nationalities, and cultures. Basically, there are two approaches to study these questions. The first one is to ask whether there are “pragmatic universals”, i.e., pragmatic regularities that are universally valid. As a candidate, we may think of the maxim of Quality (Grice 1989a), requiring that one should say the truth. Note that this maxim gets some cognitive footing when one considers the idea of Spence et  al. (2004: 1758) that truth-telling, as a routine, automated or stereotypical behavior is a “model ‘baseline’ response”, while lying is (a consciously planned, spontaneous or novel) model response that is accompanied by the suppression of the truthful response, and thus more demanding. The second approach is to empirically examine how the concept of lying is understood and how the practices of lying are acted out in different communities. Barnes (1994: 65–78) shows that there are scattered remarks in the older literature related to lying habits of others. While these remarks can easily be dismissed as prejudices, they nevertheless may be seen as indicators of the insight that cultural practices related to lying may be diverse. From a sociological point of view, Barnes distinguishes between a comparison of societies and a comparison of social classes within one single society. As for the former, he picks out evidence from studies on Greek and Lebanese populations. For instance, as Gilsenan (1976) has shown, young men or children in a Muslim Village in the Beeka valley of Lebanon are often or even typically liars: “Lies are told partly for fun, to trick one’s friends” (Barnes 1994: 70). Moreover, “lying is indulged in sometimes for

Universality of lying   

   23

its own sake, without an instrumental motive, and it is not surprising that lying is a symbol of the fantastic and unbelievable” (Barnes 1994: 70). Since lying is so pervasive, there are cue words like seriously, will you believe me, without joking, by your life and by your father’s life that guarantee the truth of what is being said. Quite obviously, the main problem with such an analysis is that a clear distinction between lying and joking/teasing is missing. If these verbal activities are examples of joking/teasing, they cannot be examples of lying since saying something that is false is not necessarily a lie. In a Greek village on the island Euboea, as studied by du Boulay (1976), lying is more functional, because it serves to strengthen family integrity by attacking the integrity of other families: “(…) but in practice ‘eavesdropping, gossiping about neighbours, inventing scurrilous explanation of events, lying to destroy another’s reputation’ were common occurrences and were accepted as legitimate strategies to follow when defending personal secrets and uncovering those of others.” (Barnes 1974: 72)

But is such a scenario typical of a certain Greek society, or typical of a closed or illiberal society? I suppose that there are several sociological factors at work that enhance a climate in which lying is normal, if not necessary. (I would like to remind movie fans of Michael Haneke’s Das weiße Band.) Barnes (1994: 77) assumes that there are also class differences with respect to lying. He suggests “that social class is yet another dimension affecting not only attitudes toward lying but also the frequency with which lies are told, the content of lies and the contexts in which they occur.” When thinking of lying in the political sphere, with regard to romantic relationships, or on the Internet, there are studies supporting the general point. However, empirical studies are scarce and require a prior definition of core concepts such as ‘social class’, ‘frequency of speech acts’, ‘content and context of a lie’. More empirical studies have been presented by Cole (1996), Brown (2002), Blum (2005), Danziger (2010), and Hardin (2010). We will focus here on Brown (2002), Danziger (2010), and Hardin (2010). Brown (2002) observes Tzeltal speakers from Tenejapa in Southern Mexico, Danziger (2010) considers Mopan Maya speakers of Southern Belize in Eastern Central America, and Hardin (2010) deals with Spanish speakers from the Eastern Region of Ecuador. As for their method, Danziger (2010) and Hardin (2010) rely on the seminal study of Coleman and Kay (1981) to be discussed in more detail in section 2.3. The hyperbolic (if not lying) title of Brown’s paper, “Everyone Has to Lie in Tzeltal” is an homage to Sacks (1975), “Everyone Has to Lie”. Both authors assume that in certain societies, lying is obligatory under certain conditions. On closer

24   

   Perspectives on lying

inspection, however, it is not so clear to me whether (at least some) speakers of Tzeltal or American English really have to lie. Sacks (1975) identifies a particular social situation in which it is necessary to lie, namely the situation in which, on the occasion of a greeting, one speaker asks the other How are you? Suppose the addressee feels lousy. Whether there follows a potential lie like Fine or an explanation of the true answer depends on the status of the questioner. It is unclear, however, whether the addressee wants to deceive the questioner, and to what extent the formula How are you? has to be interpreted as an information question or simply as a kind of greeting. Thus, as it is a kind of greeting, a reaction like How are you? would suffice; further information would be a violation of the Gricean maxim of Quantity. Brown (2002) deals at length with a category she calls “threat lies”. For instance, parents’ utterances include equivalents to The woman will ‘tzak’ you, with tzak meaning ‘grab you and take you away forever’. Besides such threats, parents and caregivers use false statements, warnings, or fake promises and they use other verbal devices such as scaring prosody in order to control the child’s behavior. But what is unclear here is whether these speech acts are lies at all. Brown (2002: 251) sees that point and tries to give an answer: “We may call them ‘control lies’, fake scary predictions (threats) or offers (bribes) by a caregiver, intended to control the child’s behaviour.” This answer is not plausible to me because it mixes up the category of sincerity with the speech acts involved. Not all utterances violating sincerity conditions are lies, as we will discuss in more detail in section 3.3. Brown (2002: 259–60) also discusses the use of lot (‘lie’) by children and finds that these uses “do not distinguish falsehoods that are intentional from those that are not, nor jokes from mistakes.” Thus it turns out that these children have only acquired a very rough concept of lying, maybe similar to the ones children have developed in other parts of the world at this age. Danziger (2010: 201) basically tries to relativize Gricean views on communication. In certain “folk models of responsibility in meaning and action” entertained in various parts of the world, Gricean “Western” views on rational communication are inadequate because “tangible outcomes of actions, including degree of damage caused” are more relevant than intentions of an actor (Danziger 2010: 201). For instance, according to an elicitation study with 15 Mopan speakers with regard to a story with an objectively false utterance which the (deceiving) speaker believes to be true, 13 speakers said that “the utterance could not longer be considered tus [the Mopan word for lie, JM], regardless of speaker’s belief state at the moment of utterance” (Danziger 2010: 209). Hence, these speakers resemble children who rely on factual falsity without considering the speaker’s belief states. However, it seems to me that results such as these are not really apt to weaken Grice’s position (cf. Danziger 2010: 200–203). Grice does not state that there is

Lying and deceiving   

   25

active intention-seeking on the part of the hearer in every conversation; instead, uncovering intentions of the speaker is a part of the general theory of understanding someone. And I would suspect that this is also the case in the Mopan society. The conditions of social life, in which the particulars of telling the truth vis-à-vis lying are evaluated according to cultural models, might be different of course.

1.7 Lying and deceiving We have seen that there is no such thing as a clear and unique indicator of lying. While it is not impossible to catch liars, as Vrij (2008: 373–418) points out, many pieces of the puzzle must be put together in order to succeed. We have also seen that the category of lying is not always clear-cut. Because of the necessity to control variables, in experiments and laboratory settings, the distinction is only made between truth and falsity. But this may be just a rough approximation to every-day situations of lying in which the boundaries of the category of lying are not always clear-cut and participants are unsure if and to what extent a certain utterance counts as a lie. Vrij (2008: 15) makes no distinction between lying and deceiving. He defines deception as “a successful or unsuccessful deliberate attempt, without forewarning, to create in another a belief which the communicator considers to be untrue.” For other researchers, the distinction between lying and deceiving is important (Mahon 2008). In most approaches, lying is just a subtype of deceiving, the latter being defined, for example, as “move or action (…) which has a goal of getting the hearer to make a false assumption” (Vincent and Castelfranchi 1981: 753; see also Castelfranchi and Poggi 1994, 1998, Gupta, Sakamoto, and Ortony 2012). The typology of deception proposed by Chisholm and Feehan (1977) includes eight types of deception and Vincent and Castelfranchi (1981) assume twelve types. It follows that not every act of deceiving is an instance of lying. What is, then, the main difference between lying and deceiving? Two differences immediately come into mind (cf. Mahon 2008). First, in contrast to lying deceiving may be non-verbal. Recall the famous example (originally given by Immanuel Kant) of someone who packs his luggage in order to let observers draw the conclusion that they are leaving. Nothing is said but it is nevertheless a case of intended deception (Chisholm and Feehan 1977: 149). Second, if verbal, deceiving is not bound to assertions. Thus, if someone asks Where has Peter gone? in order to get the hearer to assume that Peter had been there (while knowing that this was not the case), then this is a case of deception, not of lying. Accordingly, Mahon (2008: 14) points out: “It is possible for a person to deceive

26   

   Perspectives on lying

without uttering (etc.) a declarative sentence, by cursing, making an interjection or an exclamation, issuing a command or exhortation, asking a question, saying ‘Hello’ and so forth.” We will have a closer look at some of these cases in chapter 2. Let us now focus on the taxonomies of deception proposed by Vincent and Castelfranchi (1981), and Chisholm and Feehan (1977). We start with the former approach. Vincent and Castelfranchi (1981) distinguish 12 types of deception: Pretending or faking. Following the definition of Vincent and Castelfranchi (1981:  754), “this is an intentionally deceptive move obtained through counterfeiting that which the hearer is intended to assume.” Limping in order to make believe that one suffers from lameness or that one has hurt one’s leg is a standard example. Acting. Acting is closely related to pretending. For Vincent and Castelfranchi, acting is not a deceiving act per se, but pretending to act, or joking, is seen as deceptive. Examples will be given below. Lying. While pretending is non-communicative, lying is communicative. Vincent and Castelfranchi (1981: 756) argue that “the term is strictly reserved for those cases when A explicitly gives false (i.e., different from that which he assents to) information to B, and does not apply when, for example, true information is concealed by A from B.” In the parlance of Chisholm and Feehan (1977), lying is a positive deception simpliciter. For Vincent and Castelfranchi, lying is an act of double-deception: For example, by saying I love you while in fact the speaker does not love the hearer, the speaker has the deceptive goal to make the hearer believe that he loves him/her. Moreover, because the speaker acts as if he loves the hearer, he is pretending. But only lying is, according to Vincent and Castelfranchi, insincere; pretending is not. Insincerity is defined as “intentional deception regarding one’s emotions and/or assumptions and/or goals” (Vincent and Castelfranchi 1981: 757). Indirect Lying. Vincent and Castelfranchi (1981: 758) call indirect lies “that sub-set of deceptive moves whose first communicative goal may be either truthful or deceptive, but at least one of whose super-goals is deceptive and achieved by communicative means.” A case in point is a scenario, “where a speaker asserts or directly expresses a proposition he assents to while having the super-goal that B infer from it a further proposition which A does not assent to, or actively dissents from.” This is the classical scenario of lying while saying the truth.

Lying and deceiving   

   27

As Grice (1989a: 37) explains, “anyone who uses a sentence of the form X is meeting a woman this evening would normally implicate that the person to be met was someone other than X’s wife, mother, sister, or perhaps even close platonic friend.” However, if the speaker knew that the respective woman was X’s sister indeed, he could be said to be lying because his implicature was false while his assertion was true. We will come back to this issue in section 4.5. Insinuation. This type of deceiving is defined as follows: “Insinuation is a communicative act with a communicative goal which consists of an implied assumption deliberately involving a negative value-judgment of the relevant referent, fact, or effect on the hearer.” (Vincent and Castelfranchi 1981: 760) Two examples are provided: The first example draws on stereotypical meaning, e.g., when one drops the remark that someone is a Sicilian in order to convey that the respective person is a greasy Mafioso. Here, clearly a negativevalue judgment is intended. However, there may be cases in which amelioration is aimed at. The second example shows such a case. Here, an 11-year-old boy on his first day at his new school utters to his desk-mate My dad works for the B.B.C., thereby trying to convey that his father is something glamorous such as a reporter or a cameraman. In reality, however, the boy’s father is a cleaner at the studios. Thus, the desk-mate is led into an “inference to stereotype”. Reticence. This is quite a complex act of deception, since it comprises, according to Vincent and Castelfranchi (1981: 761), the following components: – “the speaker lets the hearer understand that he is keeping quiet about something” – “he makes it clear that he is keeping quiet with a particular goal in mind” – “he makes it clear what this goal is” – “he makes it clear that he has the goal of letting the hearer understand that he is leaving something unsaid” – “he lets the reasons for doing so be understood” – “he lets the hearer suspect exactly what this something is (helped by the context)” – “he lets the hearer understand that he wants him to understand what it is he is leaving unsaid” – moreover, the inferable information is typically negative or hurtful for the hearer. If this information is intentionally false, so Vincent and Castelfranchi argue, then this behavior is a case of indirect lying. If the hearer already assumes this information, we would face a “lying allusion.” The problem with reticence is, of

28   

   Perspectives on lying

course, whether the reticence of some piece of information still allows to infer the left-out information. After all, in standard cases of indirect lying, the speaker did say something while in the case of reticence, they did not say something. Half-truths. Half-truths (or half-lies) are misleading speech acts: “A speaker gives part of the truth while concealing another part of it.” (Vincent and Castelfranchi 1981: 762) The concealed part of the truth is most relevant to the hearer. Thus, s/he is led to false assumptions with regard to the whole truth. Here is an example given by Vincent and Castelfranchi: When asked about beating of his little sister, the child admits that he aimed a blow at her. This is literally true even when, in fact, the child used several punches and kicks. If the child said I aimed a blow at her, this will be interpreted as ‘I gave exactly one blow to her and no more’. The child, however, violates the maxim of Quantity that requires saying “as much as he can”. However, it is not altogether clear how much information must be given. The clearest cases are the ones where there is a certain charge against someone at stake. Only in relation to this charge and the obligation to give the complete information relevant to the case one may speak of a half-truth. Precondition or presupposition faking. According to standard wisdom, presuppositions have the status as propositions that are related to presupposition triggers and remain constant under negation. For example, in Ken realized/did not realize that the car was stolen, the presupposition that the car was stolen is active. This is not the case with a non-factive verb like to assert. Whether false (semantic) presuppositions are to be included into a general theory of lying remains to be discussed in a later chapter (see section 4.6). Vincent and Castelfranchi (1981: 763) think of precondition or presupposition faking “as a sub-type of pretending, or behaving-as-if, for it involves doing an act or making an utterance which would require certain preconditions for its appropriate use; when, however, these conditions do not exist, the speaker knows they do not exist, and his goal is that the hearer assumes that they do.” Look at the following scenario. At the speaker’s daughter’s wedding breakfast, he remarks: It is rather a pity that Anne and Mark had to be at Burleigh this weekend. Here, the speaker may very well be liable of multiple presupposition fakings: “For instance, that the speaker knows the two illustrious personages in question well enough to be on first-name terms with them, that the speaker actually invited them to the wedding, that A. and M. would have come but just could not avoid the very important business that took them elsewhere, etc.” (Vincent and Castelfranchi 1981: 763)

Lying and deceiving   

   29

Deliberate ambiguity. This type of deception is defined as follows (Vincent and Castelfranchi 1981: 763): “Given an utterance with two possible interpretations or readings in a given context, one of which is true for A and one of which is false, A may exploit the ambiguity hoping and intending that B understands the false reading.” An example from commercial advertising is No heat costs less than oil heat (the example stems from Dwight Bolinger 1973, and is taken up by Vincent and Castelfranchi 1981: 763–4). The two meanings involved are: (i) Not using any heat is cheaper than using one, (ii) oil heat is cheaper than any other heating system. Whereas (i) is trivially true, (ii) is unproven and possibly false. I think, however, that this is not a good example because it seems clear that in the context of advertizing, the second reading is the intended reading, be it true or false. A better and simpler example is the utterance He went to the bank in order to make the hearer think that he went to a river bank while really he went to the credit institute. Obfuscation. According to Vincent and Castelfranchi (1981: 764), this is a type of deception that comes close to “verbal terrorism” whereby a verbal behavior is understood that involves the speaker “assuming and intending that the hearer will not be able to understand the full import of his utterance.” This is often the case with technical jargon, or with the speech of governments, advertisers, etc. Obviously, the Manner maxim “Be perspicuous!” is violated in these cases. We will later think about this type in connection with the notion of bullshit, as developed by Frankfurt (2005) (see section 5.6). Pretending to lie. It is also possible to deceive someone by pretending to lie: “If A says what he believes to be the truth with the intention that B believe he is lying to him then he is doing what we shall call pretending to lie.” (Vincent and Castelfranchi 1981: 764) For instance, if a burglar who is suspected to be a liar is somehow caught by the police he may very well pretend to lie while saying the truth. Asked by the police where his partner has gone he may utter the truth, while the police will not believe him. Pretending to act, or joke. In this case, “the agent or speaker passes off (by some manipulation of the context) something which he believes to be true in the real world, as if it were acting, joking […] as if, in other words, what he is communicating is not true in the real world.” (Vincent and Castelfranchi 1981: 766) One important difference between pretending to lie on the one hand and pretending to act on the other is, that in the former case, there is no disloyalty between the actors because B does not trust A anyway, whereas in the latter case, there is disloyalty because B trusts A in principle.

30   

   Perspectives on lying

In comparison to the threefold distinction of Saari and Lewis in Table 1, the approach of Vincent and Castelfranchi is much more differentiated. Basically, it draws a distinction between lying on the one hand, and pretending on the other hand: Table 2: Taxonomy of deceptive acts (Vincent and Castelfranchi 1981)

deceiving lying

direct lying

pretending/faking

indirect lying indirect lying insinuation reticence half-truth deliberate ambiguity obfuscation

precondition or presupposition faking pretending to lie pretending to act, or joke

Acts of pretending/faking do not include lying, understood as essentially verbal acts, though they are nevertheless deceptive. Note that the type “precondition or presupposition faking” may be considered as a type of indirect lying. This depends on the notion of presupposition that is used here. All types of indirect lying – at least this is what I suspect – have to do with implicatures – an observation that will be worked out more clearly during the upcoming chapters. It follows that we have to draw a precise distinction between direct lying and indirect lying. Let us now have a look at the taxonomy of Chisholm and Feehan (1977). They make a difference between cases of commission – an agent L contributes causally to the belief of an agent D – and cases of omission – an agent L “has failed to do something he could have done with respect to D and the belief that p.” (Chisholm and Feehan 1977: 144) Note that in all cases p is considered as false. The authors draw a further distinction between positive deception and negative deception. Positive deception is connected to the adding or retention of a belief. Negative deception has to do with the losing of a true belief (i.e., the belief in not-p).

Lying and deceiving   

   31

Moreover, they distinguish between deception “simpliciter” and deception “secundum quid”. The notion of “simpliciter” relates to the fact that “L brought about the change from the state of not being deceived with respect to p to that of being deceived with respect to p”, as the authors explain with respect to case (a) (Chisholm and Feehan 1977: 144). The notion of “secundum quid” is related to the fact that “although L did not cause D to acquire the false belief in p, nevertheless, but for the act of L, D would not have lost or given up the belief in p.” (op. cit.) It is obvious that this is a very general taxonomy, a taxonomy that is unspecific about the particular way a commissive or omissive behavior is performed. We will not go into the details here. What is interesting to us is the status of lying, understood as a verbal act. This is the case (a), i.e., a positive deception simpliciter. In lying to D, L contributes causally to D’s acquiring the belief that p. Mahon (2008: 14) points out that some scholars reserve the term to mislead for those cases where an actor inadvertently or mistakenly deceives another person (e.g., Barnes 1997). Others argue that deceiving may be inadvertent or mistaken (Chisholm and Feehan 1977, Adler 1997). I will assume that deception always includes intention and hence I will refer to inadvertently or mistakenly deceiving as ‘misleading’ (see also Saul 2012). Table 3: Types of deception: Commission vs. omission (Chisholm and Feehan 1977)

Commission (a) Positive deception simpliciter “L contributes causally toward D’s acquiring the belief in p.” (b) Positive deception secundum quid “L contributes causally toward D’s continuing in the belief in p.” (c) Negative deception simpliciter “L contributes causally toward D’s ceasing to believe in not-p.” (d) Negative deception secundum quid “L contributes causally toward preventing D from acquiring the belief in not-p.” Omission (e) Positive deception simpliciter “L allows D to acquire the belief in p.” (f) Positive deception secundum quid “L allows D to continue in the belief in p.” (g) Negative deception simpliciter “L allows D to cease to have the belief in not-p.” (h) Negative deception secundum quid “L allows D to continue without the belief in not-p.”

32   

   Perspectives on lying

Let us take stock: A linguistic theory of lying should focus on verbal acts, i.e., such acts of lying that are conveyed by an utterance of the speaker. It should cover all the cases of lying, i.e., what Vincent and Castelfranchi (1981) call direct and indirect lying. Moreover, it should encompass the case of falsely presupposing what Vincent and Castelfranchi subsume under the category of ‘pretending/ faking’.

2 Approaching a linguistic concept of lying 2.1 Introduction In this chapter, I will outline a linguistic frame for the analysis of lying. Put simply, linguistics comprises several components (also called levels or modules) interacting with each other. These components are phonology, morphology, syntax, semantics, and pragmatics. While we focus on the semantics-pragmatics interface which is also the object of philosophy of language, it is argued that the other linguistic components also play a role in the description and explanation of lying. These components, however, are often neglected in typical approaches of philosophers of language. Moreover, even pragmatics is often reduced to an area of research where implicatures are the most important objects under investigation. Even speech acts, another classical topic in the philosophy of language, are far too often neglected. So our first task is to develop a frame of reference that is suitable for the purposes of a linguistic analysis of lying. After going into the recent debate about the semantics-pragmatics distinction, I will discuss two essential linguistic approaches. The first approach is connected with the lexical semantics of the verb to lie. Coleman and Kay’s (1981) seminal approach, in the spirit of prototype semantics, was to explore the meaning of the English verb to lie from the perspective of native speakers. Coleman and Kay (1981) argued that the meaning of the word lie “consists in a cognitive prototype to which real or imagined events may correspond in varying degrees” (Coleman and Kay 1981: 26). Though they were not primarily interested in the phenomenon of lying, arguing for prototype semantics and against standard “checklist theories” of word meaning, their approach still serves as a model for linguistic research into lying. The second approach is connected with the sentence types that are used when lying. Usually it is presupposed in the literature that lying happens by uttering one or more declarative sentences. But this assumption is far from being trivial, as I will show. In fact, sentence types and sentence modality are important structural elements of natural language, and they have a not yet completely understood relation to illocutions. Note that there are scattered remarks in the literature like in Vincent and Castelfranchi (1981: 757) who state: “Imperative utterances can be insincere, although they cannot be used as lies.” But a thorough explanation for this and similar observations is still lacking, so there is much ground to cover.

34   

   Approaching a linguistic concept of lying

2.2 The semantics-pragmatics distinction Semantics and pragmatics are both disciplines that deal with meaning. So how do they differ? This is a fundamental question in modern linguistics. In recent years, this question has been lively debated (cf. Bianchi 2004, Szabó 2005, Allan and Jaszczolt 2012). Most researchers engaged in the debate relate their approaches to the work of Paul Grice who made the by now classical distinction between ‘what is said’ and ‘what is implicated’ (Grice 1989a). Roughly, ‘what is said’ corresponds to semantics, while ‘what is implicated’ corresponds to pragmatics. For instance, ‘what is said’ is equated with “semantic content” when Bach (2002: 22) holds that “the semantic content of a sentence S in a context C is what is said in that context”. It is worth emphasizing, however, that these notions do not necessarily exhaust the respective components. For example, speech acts are pragmatic in nature, but they are by no means implicated. Numerous sophisticated approaches to the question how the distinction between ‘what is said’ and ‘what is implicated’ should be spelled out have been put forward. One basic idea is that ‘what is said’ (semantic meaning) is contextindependent, while ‘what is implicated’ (pragmatic meaning) is determined by the context. Note that for Grice indexicals and the resolving of ambiguities belong to semantics albeit these processes build on contextual information. The main mechanism of how contextual information is fed into the derivation of implicatures is pragmatic inferencing. In the process of figuring out a certain implicature the hearer integrates contextual knowledge where necessary. Most influential was the Gricean conception of pragmatic inferencing on the basis of the Cooperative Principle and an apparatus of four so-called maxims which are governed by the Cooperative Principle. The Cooperative Principle (CP) simply demands: “Make your conversational contribution such as is required, at the stage at which it occurs, by the accepted purpose or direction of the talk exchange in which you are engaged.” (Grice 1989a: 26) It is worth noting that the CP, as well as the maxims and sub-maxims are conceived as rules of rational behavior, not as ethical norms. The Gricean architecture comprises four maxims, namely the maxims of Quantity, Quality, Relevance, and Manner. All of these maxims have been subjected to critical examination, and there have been a number of counterproposals, especially from a post-Gricean point of view (see Meibauer 2006). What makes them still attractive for linguistic analysis is that they are designed for capturing pragmatic properties of phenomena like tautology, scalar implicature (Quantity), irony, metaphor, sarcasm, and belief implicature in assertions (Quality), implicatures due to thematic switch and bridging (Relevance), and implicatures due to obscurity and conjunction buttressing (Manner). These domains of applications may be subclassified according to the

The semantics-pragmatics distinction   

   35

criterion whether the pragmatic inferencing builds upon the exploitation or the observation of the CP and its maxims. Thus, tautology, irony, metaphor, sarcasm, implicatures due to thematic switch, and obscurity are based on the exploitation of maxims, while the remaining phenomena are based on the observation of maxims. Grice (1989a: 30–31) famously proposed a general scheme for the working out of a conversational implicature, thus implementing the CP and the maxims as an integral part of pragmatic or “Gricean” inferencing. This general scheme is an important contribution to pragmatic analysis and an implicit background for recent studies in experimental pragmatics. The following table shows the terminological network of Grice (1989a) with respect to “speaker meaning”: Table 1: What is said – what is implicated (Grice 1989a)

speaker meaning what is said

what is implicated

conventionally

generalized (GCI)

conversationally

particularized (PCI)

The most widely accepted type of implicature is the ‘conversational implicature’. According to Grice (1989a), there are two forms of it: Generalized Conversational Implicature (GCI) and Particularized Conversational Implicature (PCI). The following example from Levinson (2000: 16–17) nicely illustrates this distinction: (1)

The distinction between PCI and GCI (Levinson 2000) Context 1

Speaker A: What time is it? Speaker B: Some of the guests are already leaving. PCI: ‘It must be late.’ GCI: ‘Not all of the guests are already leaving.’

Context 2

Speaker A: Where’s John? Speaker B: Some of the guests are already leaving. PCI: ‘Perhaps John has already left.’ GCI: ‘Not all of the guests are already leaving.’

36   

   Approaching a linguistic concept of lying

Because the conversational implicature (‘… not all …’) triggered by some arises in both contexts, it is relatively context-independent. Relative context-independence is the most prominent property of GCIs. In addition, GCIs are normally associated with certain linguistic forms. For example, if someone utters Peter is meeting a woman this evening it is, because of the indefinite article, automatically implicated that the woman is not his wife, a close relative, etc. (cf. Grice 1989a: 37). In contrast to GCIs, PCIs are highly context-dependent, and they are not consistently associated with any linguistic forms. Note that the aptness of this distinction is denied by Relevance theory (RT). For Relevance theorists, implicatures never come as a quasi-automatic inference, cancelled if the context demands this; instead, all implicatures are steered by the main RT-principles, the First (cognitive) and the Second (communicative) Principle of Relevance (Carston 2002, 2010). The distinction between ‘conversational’ implicatures and ‘conventional’ implicatures draws on the observation that in coordinations like Anna is rich but she is happy the truth conditions are the same truth conditions as those of the coordination Anna is rich and she is happy, with the exception of the contrastive meaning of but. This meaning is not truth-functional, and it is not context-dependent either; hence, there is some motivation for assuming the category of conventional implicature. There are harsh opponents of this approach (Bach 1999), but there are also attempts at reviving the notion (Potts 2005, Horn 2010). While Bach (1999) argues that sentences may have more than one proposition, and there is no need of invoking a “mythical” notion of conventional implicature (cf. Conrad 2011), Potts (2005) revives the notion, drawing on cases like nominal appositives, expressives like damn etc. which do not attribute to what he calls at-issue content (cf. Kaplan 1999, McCready 2010, Potts 2012). The rough picture of the major camps involved in the debate on the proper semantics/pragmatics divide is that we have the so-called ‘Neogriceans’ on the one hand, and the ‘Relevance theorists’ on the other. Neogriceans by and large tend to defend the conceptual value of Gricean maxims or principles (e.g., Levinson 2000, Atlas 2005, Horn 2009), while Relevance theorists argue against such maxims and refer to general cognitive principles such as the Principles of Relevance (e.g., Sperber and Wilson 1995, Carston 2002). Indeed, there are many differences in the pragmatic architecture of the rivalling camps’ approaches as well as in the coverage of empirical phenomena. But, quite surprisingly, Neogriceans as well as Relevance theorists go for the assumption that propositional structures are systematically underdetermined and therefore are in need of enrichment. These processes of enrichment are of an essentially pragmatic nature, and hence pragmatics is conceived of as being able to influence semantics.

The semantics-pragmatics distinction   

(2)

   37

Sources of underdeterminacy (Carston 2002, drawing on Gross 2001)

sources of underdeterminacy

examples

proposition pragmatic processes

multiple encodings

She surprised the man in the pyjama.

severable

context-dependent resolution

indexical references

Now I like it.

variable

context-dependent filling

missing constituents

– In the cave. – Super 90 is better. – This fruit is red.

incomplete adding of conceptual constituents

unspecified scope of elements, e.g. negation

– She didn’t toast the bread with a knife in the bathroom. – I didn’t eat breakfast.

full

derivation of speakerintended proposition + ?resolution of ambiguities + ?filling of variables

underspecificity or weakness of encoded conceptual content

The way was uneven.

full

derivation of speakerintended proposition + adjustment of conceptual constituents

overspecificity or narrowness of encoded conceptual content

The room was silent.

full

derivation of speakerintended proposition + adjustment of conceptual constituents

Several terminological proposals are on the market to fix the phenomenon of pragmatically steered propositional enrichment: ‘explicature’ (Carston 2002), ‘impliciture’ (Bach 1999, Garrett and Harnish 2007), ‘pragmatic intrusion’ (Levinson 2000), or ‘intuitive content’ (Recanati 2004). The detailed comparison of these proposals, or, more generally, the explicit/implicit distinction, is on the pragmatist’s research agenda. For the distinction between impliciture and explicature, see Bach (2010) and Carston (2010). Insofar as ‘underdeterminacy’ is taken as a serious linguistic phenomenon, there is a built-in tendency within the approaches mentioned to restrict the realm of truth-conditional semantics, or to downplay its importance. According to those approaches, there are truth-conditions or logical forms of sentences, but they only occur at a certain stage in the development of the comprehensive meaning of an utterance. The tendency sketched has of course provoked conjectures. For instance, some researchers defend a classical, minimalist approach to the truth conditions of a sentence, and consequently assume a more powerful apparatus for pragmatic interpretation (cf. Borg 2004, Cappelen and Lepore 2004; see also Borg 2013).

38   

   Approaching a linguistic concept of lying

‘Minimalism’ is thus opposed to ‘Contextualism’ which believes that context influences semantics, not only in the case of indexicals, but also in numerous further aspects. Recanati (2010) summarizes the minimalist picture as follows: Table 2: The minimalist picture (Recanati 2010)

sentence meaning (character) + pre-propositional pragmatics (saturation) = what the sentence says + post-propositional pragmatics (modulation, implicatures, etc.) = what the speaker means

‘Saturation’ is, according to Recanati (2010: 4), defined as “a pragmatic process of contextual value-assignment that is triggered (and made obligatory) by something in the sentence itself, namely the linguistic expression to which a value is contextually assigned”. Indexical variables that are contextually instantiated are a case in point. Recall that Grice assumed that indexical and ambiguity resolution happens within the realm of ‘what is said’. ‘Modulation’, by contrast, is a “free” pragmatic process like in the case of metaphors. Recanati’s (2010) “truth-conditional pragmatics”, as a contextualist approach, says that truth-conditional processes may be affected not only by saturation but also by modulation. However, considering further approaches, the picture becomes more complex. In her review of different positions in logical space, Borg (2010: 96) proceeds from the following example taken from Travis (1989: 18–19). It demands that you “imagine a fridge which contains only a small puddle of milk on its floor”. Then two contexts are compared (as a thought experiment, displaying a ‘context-shifting argument’ (CSA) in the sense of Cappelen and Lepore 2004; see also (1) above): (3)

Context 1 Hugo is dejectedly stirring a cup of black coffee. Noticing this Odile says There is milk in the fridge. Context 2 Hugo has been cleaning the fridge. Odile opens the fridge door and says There is milk in the fridge.

The semantics-pragmatics distinction   

   39

While the utterance of Odile was the same on both occasions, the respective truth value might be different. In the first context, what was said was false, in the second context, what was said was true. The point is that milk must be “enriched” in some way so that it denotes ‘milk in a bottle that can be used for coffee’, etc. (Note in addition a fact not mentioned by Borg, namely that in the first scenario, the utterance would be interpreted as an advice, while in the second scenario, it would be interpreted as a reproach.) Borg (2010) sketches the following five approaches to the semantics/pragmatics divide, namely Minimal semantics, Indexicalism, Contextualism, AddedParameters-Approach, and Occasionalism. (Thus, the approach taken by Saul 2012, namely to draw a distinction between “constrained conceptions”, “unconstrained conceptions”, and “austere conceptions” may turn out as far too simple.) ‘Minimal semantics’ has the point of view that (formal) semantics deals with truth conditions (of propositions) which reflect the compositionality and hence the syntactic structure of a sentence. The only syntactically relevant, contextsensitive elements are the indexicals, so that Minimal semantics is not totally context-free. Aspects of meaning that are not captured by truth conditions are handled by pragmatics. Note that Grice (1989a) is a minimal semanticist according to this sketch. Others include Borg (2004) and Cappelen and Lepore (2004). ‘Indexicalism’ differs from Minimal semantics “by expanding the amount of syntactically licensed context-sensitivity in our language” (Borg 2010: 98–99). Thus, it might be argued that there are further “hidden” indexicals over and above the usual ones, or that our understanding of what an indexical is must be widened. Such positions are ascribed to Stanley (2000, 2005a, b). ‘Contextualism’ has the point of view that pragmatic principles are active in the determination of truth conditions. Thus it is denied “that the route to semantic proponents runs exclusively via syntactic content” (Borg 2010: 99–100). Proponents of such an approach are researchers like Levinson (2000), Sperber and Wilson (1995), Carston (2002), Jaszczolt (2005), Recanati (2010). The ‘Added-Parameters-Approach’ “treats contextual material (in at least some cases) as contributing to the parameters against which a proposition is assessed for truth or falsity” (Borg 2010: 100). For instance, with respect to context-sensitive predicates like rich or tall, one could assume that there are different contextual parameters added to the truth-evaluation of sentences like Jill is rich and Jake is tall: Depending on these parameters, Jake is tall may be true on one occasion, and false on another occasion. Thus, ‘Nonindexical Contextualism’ holds that such sentences do not have context-invariant propositions (MacFarlane 2009). Finally, ‘Occasionalism’ rejects the “idea of ‘filling out’ a linguistically expressed content to get to the complete content the speaker had in mind, for this

40   

   Approaching a linguistic concept of lying

kind of context-independent content is non-existent” (Borg 2010: 103). Occasionalism has the radical point of view (in the tradition of the late Wittgenstein) that there is no such thing as a context-independent meaning. All meanings are intimately related to contexts because they arise only in contexts. Even if utterances were enriched (as in the case of missing constituents), the sort of enrichment (say, the underlined material in I didn’t eat breakfast this morning) is in need of contextual interpretation, too (cf. Travis 1989). Discussing weaknesses and strengths of these approaches, Borg (2010: 103) suggests that “it is really the old warhorses of formal semantics and use-based theories of meaning which remain standing”. I would like to add that with the advent of experimental pragmatics, some of the typical battles of the “border wars” (Laurence Horn) will be enriched with more empirical substance (see for instance Doran et al. 2012). Throughout this book, I will take a minimalist model such as the one presented in Table 2 as my frame of reference. What cannot be achieved in this book is to find the “right” model of the semantics-pragmatics interface. My goal is to analyze lying with respect to the minimalist model of the semantics-pragmatics interface, and to point out where this model is in need for (linguistic) enrichment. This includes contextualist perspectives, especially with regard to the typical issues coming up in the debate between minimalists and contextualists: whether there is a difference between the said and the implicated, between GCI and PCI, between implicature and explicature/impliciture, between conversational implicature and conventional implicature, and what notion of context is needed. Looking again at the model in Table 2, we may assume that the notions of truth and truth condition is contained in the notion of ‘what is said’. Truth is of course an important notion in the analysis of lying. What is lacking in this model is a clear relation between ‘what is said’ and ‘what is implicated’ to illocutions (speech-act types). However, we need such a relation in order to analyze lying. Moreover, ‘what is said’ and illocutions are dependent on a proper notion of sentence type, but we do not know how sentence types fit into the above mentioned minimalist model. And last but not least, there is a relation between sentence types and illocutions, an issue that has traditionally been discussed under the label of ‘speech-act assignment’. Hence, the above model has to be fleshed out in a number of perspectives. Before we turn to the connection between sentence type and the act of lying, we will shortly go into another linguistic perspective on lying, one that focuses on the lexical meaning of the verb to lie.

Lexical semantics: The meaning of to lie   

   41

2.3 Lexical semantics: The meaning of to lie One important observation of speech-act theoreticians like Searle (1969) was that there is a connection between the act of promising and the meaning of the verb to promise. Similarly, there must be a relation between the act of lying and the meaning of to lie. It goes without saying that speakers of a language should be aware of the meanings of the lexical items of that language. German speakers, for example, should know the meaning of lügen (‘to lie, to prevaricate’) and the difference in meaning of the prefixed variants anlügen, belügen, where the prefix triggers a change in the verb’s argument structure. Moreover, they should be able to judge whether related verbs such as schwindeln and flunkern, as well as their prefixed variants such as anschwindeln, beschwindeln, anflunkern, vorflunkern etc., and certain idioms like lügen wie gedruckt and lügen, dass sich die Balken biegen are synonymous or not. For instance, there is a ranking lügen > schwindeln > flunkern with “>” meaning ‘stronger than’. According to Harras et al. (2004: 45), the user of schwindeln conveys that they do not evaluate a lie as grave or far-reaching (“(…) gibt damit zu verstehen, dass dieser die geäußerte Lüge als nicht sehr schwerwiegend oder konsequenzenreich für den Hörer H der Bezugssituation einschätzt.”) With flunkern, this effect of downplaying the lie is even stronger so that flunkern is often used when parents talk to their children about their lies. From a lexicographic as well as from a speech-act theoretical point of view, the meaning of lügen has to be related to a definition of lying. Harras et  al. (2004: 43) paraphrase lügen with “etwas sagen, was man selbst für falsch hält” (‘to say something which one believes to be false’), or, more elaborate (S = speaker, H = hearer, Sa = speech act, P = proposition): (4)

Ein Sprecher S äußert (einem Hörer H gegenüber) eine oder mehrere Sa(P), um zu bewirken, dass H erkennt, dass S P für wahr hält; S hält P nicht für wahr. ‘A speaker S utters (to a hearer H) one or more Sa(P) in order to make H recognize that S considers P to be true; S does not consider P to be true.’

However, one may ask whether definitions such as (4) are correct definitions from the perspective of the everyday speaker. Basically, Coleman and Kay’s approach is an attack against so-called checklist semantics. Checklist semantics holds that “the definition of a semantically complex word […] consists of a set of features and components such that a given object (physical or otherwise) is aptly labeled by the word just when it possesses the property named by each feature in the definition.” (Coleman and Kay 1981: 26)

42   

   Approaching a linguistic concept of lying

For instance, if bachelor entails the features [+human], [+male], [+ unmarried], then “the list of features amounts to a set of necessary and sufficient conditions which a thing must satisfy in order to be an instance of the category labelled by the word” (Coleman and Kay 1981: 26–27). If a given man is not unmarried, then the referent of man does not belong to the category ‘bachelor’, and thus the word bachelor is not applicable to that specific man. The prototype view defended by Coleman and Kay argues “that semantic categories frequently have blurry edges and allow degrees of membership”; in addition, “applicability of a word to a thing is in general not a matter of ‘yes or no’, but rather of ‘more or less’” (Coleman and Kay 1981: 27). Thus, they argue that the meaning of the verb to lie supports such a prototype view of word semantics. They assume that the prototypical lie is an utterance in which the speaker S asserts some proposition P to an addressee A, and which comprises the following further conditions: (5)

a.

P is false.

b.

S believes P to be false.

c.

In uttering P, S intends to deceive A.

Thus lying is characterized as a deliberate, false, and deceptive utterance. In order to verify their hypothesis, Coleman and Kay ran an experiment. Eight stories were presented to 71 people who had to decide on a 7-point-scale whether an utterance contained within the respective story was a lie. The stories were construed in such a way that there was a variation of the conditions in (5). The respective variation is shown in square brackets; e.g., in story III. (Pigfat), P is false (+), S believes P is false (–), S intends to deceive (+). Moreover, the total scores on the basis of the seven-point-scale and the mean scale scores are added.¹ (6)

The eight stories in Coleman and Kay (1981)

I. (Moe) [+, +, +] 466; 6.96 Moe has eaten the cake Juliet was intending to serve to company. Juliet asks Moe, ‘Did you eat the cake?’ Moe says, ‘No.’ Did Moe lie?

1 The maximal score was 496 (= scale score of 7 x 67 subjects). The minimum possible score was 67. The mean scale score equals the total score divided by 67.

Lexical semantics: The meaning of to lie   

   43

II. (John) [–, –, –] 71; 1.06 Dick, John, and H.R. are playing golf. H.R. steps on Dick’s ball. When Dick arrives and sees his ball smashed into the turf, he says, ‘John, did you step on my ball?’ John replies, ‘No, H.R. did it.’ Did John lie? III. (Pigfat) [+, –, +] 245; 3.66 Pigfat believes he has to pass the candy store to get to the pool hall but he is wrong about this because the candy store has moved. Pigfat’s mother doesn’t approve of pool. As he is going out the door intending to go to the pool hall, Pigfat’s mother asks him were he is going. He says, ‘I am going by the candy store.’ Did Pigfat lie? IV. (Katerina) [–, +, +] 346; 5.16 One morning Katerina has an arithmetic test she hasn’t studied for, and so she doesn’t want to go to school. She says to her mother, ‘I’m sick.’ Her mother takes her temperature, and it turns out to Katerina’s surprise that she is really sick, later that day developing the measles. Did Katerina lie? V. (Schmallowitz) [+, +, –] 315; 4.70 Schmallowitz is invited to dinner at his boss’s house. After a dismal evening enjoyed by no one, Schmallowitz says to his hostess, ‘Thanks, it was a terrific party.’ Schmallowitz doesn’t believe it was a terrific party, and he really isn’t trying to convince anyone he had a good time, but is just concerned to say something nice to his boss’s wife, regardless of the fact that he doesn’t expect her to believe it. Did Schmallowitz lie? VI. (Mary) [–, –, +] 233; 3.48 John and Mary have recently started going together. Valentino is Mary’s ex-boyfriend. One evening John asks Mary, ‘Have you seen Valentino this week?’ Mary answers, ‘Valentino’s been sick with mononucleosis for the past two weeks.’ Valentino has in fact been sick with mononucleosis for the past two weeks, but it is also the case that Mary had a date with Valentino the night before. Did Mary lie? VII. (Nurse Braine) [+, –, –] 199; 2.97 Two patients are waiting to be wheeled into the operating room. The doctor points to one and says, ‘Is Jones here the appendectomy or the tonsillectomy?’ Nurse Braine has just read the charts. Although she is anxious to keep her job, she has nevertheless confused the charts in her mind and replies, ‘The appendectomy,’ when in fact poor Jones is the one scheduled for tonsillectomy. Did Nurse Brain lie?

44   

   Approaching a linguistic concept of lying

VIII. (Superfan) [–, +, –] 309; 4.61 Superfan has got tickets for the championship game and is very proud of them. He shows them to his boss, who says, ‘Listen, Superfan, any day you don’t come to work, you better have a better excuse than that.’ Superfan says, ‘I will.’ On the day of the game, Superfan calls in and says, ‘I can’t come to work today, Boss, because I’m sick.’ Ironically, Superfan doesn’t get to go to the game because the slight stomach ache he felt on arising turns out to be ptomaine poisoning. So Superfan was really sick when he said he was. Did Superfan lie? The hierarchy proposed by the subjects was as follows: The Moe-story displayed a lie to a high degree, while the John-story was considered as a lie to a low degree. All in all, the stories with utterances containing more prototypical elements were rated higher than the stories containing utterances with less prototypical elements. The exception was story VIII that scored higher than story III. We will later come back to a possible reason for this specific result. (7)

1.

I

(Moe)

+++

6.96

2.

IV

(Katerina)

++–

5.16

3.

V

(Schmallowitz)

–++

4.70

4.

VIII

(Superfan)

–+–

4.61

5.

III

(Pigfat)

+ – +

3.66

6.

VI

(Mary)

+ – –

3.48

7.

VII

(Nurse Braine)

– – +

2.97

8.

II

(John)

– – –

1.06

A further important result is that the conditions (5a–c) can be ranked according to their importance. For instance, a comparison between story III and story V shows the following properties: (8)

A comparison between Story III and Story V

Story III P is false S does not believe P is (Pigfat) false Story V P is false S does believe P is false (Schm.)

S intends to deceive

3.66

S does not intend to deceive

4.70

Lexical semantics: The meaning of to lie   

   45

This comparison reveals that element (5b), ‘false belief’, is more important than element (5c), ‘intent to deceive’, for story V receives a higher mean score than story III. The comparison of such pairs of stories differing in exactly two elements yields a unique ordering of importance: believe to be false (5b) > intends to deceive (5c) > false in fact (5a), with “>” meaning ‘more important than’. The results of Coleman and Kay’s study appear to be straightforward. However, some criticism has been put forward. Thus, Tsohatzidis (1990) argues that Coleman and Kay’s approach is faced with problems concerning “clear” cases as well as “unclear” cases. The first example goes like this (Tsohatzidis 1990: 441): (9)

In addressing a class which includes Paul and Maria, a teacher asks, ‘Who wants to answer my next question?’ Maria has not given to anyone in the class any evidence of any sort concerning her willingness or unwillingness to answer the next question, but Paul, whose only concern is to avoid having to answer himself, immediately responds, ‘Maria wants to.’

Tsohatzidis (1990: 441) argues that  – despite the applicability of Coleman and Kay’s three prototypical criteria – we would not say that Paul is a liar because “Maria would be the only person in the class who could possibly lie or not lie about her own desires”. I do not agree. I think that Paul’s utterance has to be understood as “I have evidence that Maria wants to answer the teacher’s question”, and this qualifies for a lie. While (9) is intended to show that the conditions are fulfilled but no lie has been performed, the next example is intended to show the converse case – the conditions are not fulfilled, but a lie has been performed. The second example is given in (10) (Tsohatzidis 1990: 441): (10)

Paul meets Maria one day after her serious accident, and says to her, ‘I am very sorry about your accident’, while being obviously unable to resist laughing.

Because of the obvious contradiction between Paul’s verbal expression on the one hand, and his facial manifestations of joy on the other, Maria may call him a liar although it might be the case that Paul does not fulfil all criteria for lying: he was in fact sorry, he believed himself to be sorry, and he did not intend to deceive Maria. The problem with this example is the status of the facial expression as an indicator for insincerity. In a broad sense, people who show insincere behavior may be called liars. E.g., someone who breaks a promise may be called a liar.

46   

   Approaching a linguistic concept of lying

However, a definition of the speech act of lying (without any interpretation of accompanying gestures or mimics) cannot cover cases like (10). Tsohatzidis (1990) goes on to tackle two stories of the original story set, pointing out how the subjects’ preference of the intermediate regions of the sevenpoint-scale may have to do with the construal or content of the stories. Have a look at story VI (Mary). Here, it can be predicted that some subjects are unsure about how to judge because Mary’s answer may be interpreted on two levels. With regard to the literal meaning of her utterance, Mary is not lying, while with regard to the implicature of her utterance (‘I haven’t see him.’) she is lying. I agree and point to section 4.5 where I will argue that deliberately false implicatures should be conceived of as part of the lie (cf. Meibauer 2005). Thus, Tsohatzidis (1990: 442–43) argues that it is unwarranted to interpret the subjects’ ratings as evidence for the indeterminate meaning of lie. The other story Tsohatzidis considers more closely is story VII (Nurse Braine). Here, he again points to an aspect of the story that has possibly led to diverging evaluations. Nurse Brains’ behavior is best classified as a mistake. So the task to classify it as a lie or a non-lie was sort of misleading in the first place. The moral Tsohatzidis draws is that the results do not show evidence for intermediate categories, but rather “that it is unclear whether it was the ordinary use of the word lie that C and K’s experiment was really concerned with” (Tsohatzidis 1990: 444). Why had story VIII (Superfan) scored higher than story III (Pigfat)? An answer that draws on methodological aspects is given by Coleman and Kay (1981: 38–39): Participants were, according to their interpretation of the results, unsure (a) “whether Superfan was calling in sick because of his slight stomach ache, or because he wanted to go to the game”, and (b), “whether Superfan intended his boss to deduce that he was going to the game, or was merely making an inept attempt to deceive, hoping his boss would not remember the day of the game”. Thus, Coleman and Kay intended the Superfan story to lack the element “In uttering P, S intends to deceive A”, but several participants probably judged that the story contained this element, i.e., they assumed that Superfan wanted to deceive his boss. The Nurse Braine example (VII) was also problematic from the perspective of the judges. This example was intended to represent the case of a simple honest mistake, but several judges, among them medical people, considered it as a straightforward lie. Coleman and Kay (1981: 39) comment that these judges “took something akin to a ‘performative’ analysis of VII, and classed THAT as a lie”. They propose inferencing along the following lines: “Nurse Braine said that Jones was the appendectomy patient. In saying he was, she committed herself to knowing he was. Since in fact he was not, she could not have known that he was, and therefore she lied.” That is, if someone asserts something they are unsure of, then this counts as a lie – or, in other words, if someone deliberately does not

Interface approach: Lying and sentence types   

   47

observe the maxim of Quality then this amounts to be equal to lying. Surely, the gravity of Nurse Brain’s error may force such a “performative reading”. Coleman and Kay are eager to point out some possible sources of interference in the data – the use of names like “Pigfat” (evoking an unpleasant character), the artificiality of some stories that has to do with the problem of construing a set of paradigmatically contrasting stories, and above all, the problem that people might interpret situations within the context of what they consider likely in the real world. But all in all, Coleman and Kay created a test design that could serve as a model for other similar enterprises, whether one concludes from their findings that the meaning of lie must be prototypical in nature or not. As already mentioned in section 1.6, their study has partly been replicated in a number of studies, most recently by Danziger (2010) and Hardin (2010). Participants of the latter study were 49 speakers of Spanish living in Eastern Ecuador. Hardin compares her findings to the earlier studies of Coleman and Kay (1981) with speakers of English, and Cole (1996) with speakers of Makkan Arabic, and finds a number of cultural differences (see her summarizing Table  7 with the mean scores for all stories). For instance, “the Spanish results suggest a clash of cultural values in situations where information is sought or suggested” (Hardin 2010: 3209). In her Story 9 (stranger’s asking for a certain direction) and Story 11 (asking for the delivery of cookies in the context of a food story), many Spanish participants gave false answers for social reasons. They know that these answers are false but assume that everyone knows that such answers are lies, so they do not presume an intent to deceive. Such a behavior seems culturally well-known, but contrasts with U.S. culture where giving repeatedly false information comes close to being pathological. While the prototypical approach researches the concept of lying from the point of view of a certain group of speakers, the philosophical tradition has always been aiming at erecting more or less formal models on the basis of conceptual distinctions (for instance, the ones proposed by Chisholm and Feehan). Interestingly, both approaches work with stories representing contexts (Meibauer 2012a), and both approaches succeed in identifying certain points in which speakers’ and philosophers’ intuitions vary, for instance, with respect to the ‘intent to deceive’. Both approaches also tend to neglect formal aspects of lying to which we turn in the next section.

2.4 Interface approach: Lying and sentence types Judging by the vignettes provided, Coleman and Kay (1981) seem to assume that the declarative sentence is the sentence type that is most suited for lying.

48   

   Approaching a linguistic concept of lying

Accordingly, the specific sentence type does not go into the definition of lying. Even speech-act theoretical notions like “assertion” do not play a role. From a linguistic point of view this is unsatisfying, because one does not know where to place lying in the overall grammar-pragmatics architecture of language. Convincing definitions of lying should show clear relations to grammar and pragmatics, especially with respect to the semantics-pragmatics interface. There are two fundamental observations about lying that force linguists to consider interfaces between linguistic levels: First, there is the observation that lying must be defined with respect to the semantic notion of proposition (see (5) above). Propositions need a syntactic bearer and this is the sentence type. Second, there is the observation that lying, analyzed as a specific communicative act, is always bound to specific sentence types. Quite generally, we need to model the relation between sentence types and communicative acts that may be carried out by using them. This problem has been discussed in the literature under the heading of ‘speech-act assignment’ (Gazdar 1981, Beyssade and Marandin 2006, see also Kissine 2012, 2013), but it also shows up when issues of indirect speech acts or the general connection between sentence type, sentence modality, and illocution are discussed (Reis 1999, Meibauer 2013a). Let us start with some basic observations. It is beyond every doubt that utterances based on sentences display a certain sentence type. (There are of course sub-sentential utterances that deserve a separate analysis, see Stainton  2006.) Like most speech-act theorists (e.g., Searle 1969, Harnish 1994, König and Siemund 2007), I assume that there is a fundamental correlation between sentence types on the one hand, and speech-act types (or illocutions) on the other. Somewhat simplified, we can juxtapose sentence types and speech-act types in the following manner: (11) a.

Examples for sentence types The postman rings twice.

Speech-act type Assertion

b. c.

Does the postman ring twice? Who rings twice?

Confirmative Question Wh-Question

d.

Ring twice!

Command

e.

Boy, how he rings!

Exclamation

f.

If he only rang twice!

Wish

The examples in (11) show a declarative sentence type, a yes/no interrogative sentence type, a wh-interrogative sentence type, an imperative sentence type, an exclamative sentence type, and an optative sentence type.

Interface approach: Lying and sentence types   

   49

Note that such a relation is conceived of as a direct relation. It does not account for indirect uses of the respective sentence types, e.g., when the utterance The postman rings twice is used to convey a command. Indirectness is sometimes used as an argument against the idea that there is a basic form-function fit as in (11). However, the sceptic has to explain why there is a variety of sentence types at all. In principle, there could be languages which contain only one sentence type; yet it is not plausible that we ever come across such a language. From the perspective of grammar, the question is how the grammatical make-up of a sentence type relates to speech-act types (or an illocutionary potential); from the perspective of pragmatics, the question is which sentence types are suited for conveying a certain illocution. It is obvious that there are grammatical properties which determine the respective sentence type, among them the syntactic position of the finite verb and wh-elements, verbal mood (indicative, subjunctive, etc.), certain lexical elements like complementizers, question words, particles, and prosody (cf. Altmann 1993 for a taxonomy of sentence types in German, König and Siemund 2007 for a review of sentence types in other languages). How these properties systematically interact is the object of recent research (see Meibauer, Steinbach, and Altmann 2013). It is also reasonable to assume that formal properties encode semantic properties of sentence types. Sometimes these semantic properties are called sentence mood or sentence modality. In formal semantics, sentence mood may be represented as an operator taking propositional content in its scope (see Krifka 2011). How such an operator, e.g., QUEST for questions, has to be interpreted, is a matter of much discussion. With respect to German, one finds different analyses, for instance referential approaches, assuming an open proposition with respect to wh-interrogative sentences, and attitudinal approaches, associating certain propositional attitudes with the respective sentence types, e.g., “Speaker wants to know for which x p is true”. In the following, I will ask which sentence types are suited for lying. My main aim is to argue that this is an empirical question that should be taken seriously by researchers studying lying. I will show that there are several sentence types beyond the prototypical declarative sentence that can be used by a liar. Each sentence type has its own lying potential. First of all, consider the declarative sentence type. It goes without saying that it is suited for lying. The connection between the declarative sentence and a possible lie is so strong that it is usually assumed without further discussion. We will come back to this connection when we discuss the concept of assertion in chapter 3. However, at least in German, the declarative sentence is not as uniform as it appears at first sight (cf. Oppenrieder 2013). As a matter of fact, we have not only

50   

   Approaching a linguistic concept of lying

so-called V2-declaratives in which the prefield is filled with one constituent, but also V1-declaratives in which the prefield is left empty. (12)

a.

Da steht ein Mann an der Tür. (V2, fact-related reading) There stands a man at the door. ‘A man stands at the door.’

b.

Ø

Steht

ein Mann an der Tür. (V1, event-related reading)

Interestingly enough, there is evidence that the formal property of verb position determines the suitability of a declarative sentence with regard to lying. Thus, there are German declarative sentences with the finite verb in first position (V1), roughly displaying the same meaning as their V2-equivalents. In (12), this meaning is ‘A man stands at the door’. Yet there is a difference between the two structures, since V1-declaratives as in (12b) have an essentially narrative character (therefore being preferred structures, say, in jokes) and show no structuring between topic and commentary (cf. Önnerfors 1997). Hence, it seems that because of the presentation of the propositional content as an event in the former case, and as a fact in the latter, as Reis (2000: 224) has argued, V1-declaratives are not structures that can be adequately used for lying. Only V2-declaratives allow for a proper assertion of the propositional content, and, consequently, for lying. Next, we consider interrogative sentences. These sentences are used to convey questions and they are not regularly connected with assertions. If it is supposed that lies are insincere assertions, it follows that it is impossible to lie by using interrogatives. However, it is perfectly possible to deceive others by using interrogatives. This has to do with certain presuppositions that are triggered by such sentences. Consider the following examples: (13)

a.

Does Paul like the book?

b.

Who comes to the party?

The speaker of (13a) presupposes that there is a certain book that Paul knows. But if s/he knows that there is not such a book, s/he deceives the hearer. Similarly, in (13b) it is presupposed (by way of the wh-element) that there is someone who comes to the party. But if the speaker knows that there is not such a person, s/he deceives the hearer, too. Now it could be argued that the presupposition connected with the expression the book in (13a) is not specific for the interrogative, since in a normal declarative sentence, like Paul likes the book it is presupposed, too, that there is a certain book that Paul likes. I agree. But note that the crucial point with confirma-

Interface approach: Lying and sentence types   

   51

tive interrogatives is that the hearer should answer with yes or no. Thus, the presupposition that there is a certain book appears to be more backgrounded than in the corresponding declarative Paul likes the book in which one might immediately respond with Which book (do you mean)? There is a certain dispute in the literature whether the existential presupposition in (13b) is a real presupposition at all. The main argument is that is does not survive the negation test: (14)

Who does come to the party? Who doesn’t come to the party?

 >> ‘Someone comes to the party.’ *>> ‘Someone comes to the party.’

In standard cases, the presupposition is constant under negation. Therefore, it has been argued that the existential presupposition in wh-interrogatives is in fact a conversational implicature. (Since I will argue later that intentionally false implicatures should be included in a comprehensive definition of lying, it would be consequent to include the case of wh-implicatures into that definition. I will come back to this question in section 4.5. Note in addition that deliberately false presuppositions may be included into a general definition of lying, similar to my argumentation with respect to deliberately false implicatures. See the detailed discussion in section 4.6.) Let us now turn to assertive questions, i.e., declarative sentences with a rising intonation contour (cf. Poschmann 2009, Reis 2012): (15)

a.

Es It ‘It rains?’

b.

Regnet es? Rains it? ‘Does it rain?’ Das Geschäft The shop ‘The shop is closed?’

c.

d.

regnet? rains?

Ist das Is the ‘Is the shop closed?’

ist is

geschlossen? closed?

Geschäft shop

geschlossen? closed?

There is a long-standing debate about the question how these constructions are to be integrated into the system of sentence types and sentence mood. Basically, one

52   

   Approaching a linguistic concept of lying

can argue that they are declarative sentences because of the V2-structure, that they are interrogative sentences because of the rising intonation contour, or that they are a sort of “mixed type”, as Altmann (1993) has put it, because properties of the prototypical declarative and (confirmative) interrogative sentence are mixed (cf. Gunlogson 2001, Poschmann 2009). We will not go into that debate here, but simply point out that the utterances in (15) are questions, i.e., with respect to (15a), the speaker wants to know whether it is the case that it rains. What is, then, the difference to the standard confirmative question Regnet es? (‘Does it rain?’). Obviously, the difference must have to do with the speaker’s assumption that something is the case. The speaker is slightly biased in their belief. Thus, they should only ask the question (13a), if they believe to have certain evidence that it rains. The standard confirmative question, as pointed out above, appears not to be suited for lying (the problem of presupposition put apart). What about assertive questions? Do they allow for lying? Like questions, they have the same status as other questions do – because of their lack of assertive force, they are not suited for lying. However, if the speaker knows that it is not raining and asks Es regnet? (‘It rains?’), thereby suggesting that they believe that it is raining but are not totally sure about it, one could argue that they are at least deceiving the hearer about their true belief. Thus, so called assertive questions appear to display a certain deceiving potential that standard confirmative questions do not possess. It is similar with echo questions. They come in two varieties, namely yes/ no-echo interrogatives and wh-echo interrogatives. (16)

Speaker A:

Speaker B:

(17)

Speaker A:

Speaker B:

Speaker A:

Bayern hat schlecht gespielt. Bayern has badly played. ‘Bayern played badly.’ Bayern schlecht gespielt? Bayern has badly played? Bayern hat schlecht gespielt! Bayern has badly played. ‘Bayern played badly.’ Bayern hat WAS? Bayern has WHAT? ‘Bayern has WHAT?’ Schlecht gespielt. Badly played. ‘Played badly.’

(Dass ich nicht lache!) (I laugh my socks off!)

Interface approach: Lying and sentence types   

   53

The utterance of speaker B looks like an assertive question, but, other than initiative assertive questions like Regnet es?, it is a reaction to speaker A’s statement. For example, speaker B did not understand speaker A for some acoustic reason (for example, there was a noisy Lamborghini in the background), or s/he did understand it and repeats it for purely rhetorical reasons (“incredulity question”, see the utterance in brackets). It is obvious that in both cases, deception is possible: speaker B did very well understand what speaker A said, or s/he actually shared the conviction of speaker A. There is considerable discussion about echo questions in the German literature – whether wh-echo questions constitute a sentence type of their own. Reis (1991, 2012) denies this for several reasons. One main reason is that there are not only echo questions in the narrow sense, that is, reactions to a declarative sentence displaying a wh-element that is not in the prefield, but there are also echoes of confirmative interrogatives, wh-interrogatives, imperatives, exclamatives and optatives. The common feature of these utterances is the atypical position of the wh-element in the middle field. Hence, for Reis, there is no proper echo-wh-sentence type, but only a class of constructions that share the position of the wh-element as a mutual feature. In this approach, the question arises how to account for the typical interrogative force of the wh-echo question. Reis (1991) proposes that the interrogative force has the status of a conversational implicature. Her solution, basically, is that the hearer, when confronted with a wh-echo utterance, notices that the wh-element is in an atypical position, and therefore, via Gricean maxims, concludes that it is a question that is posed. I cannot go into the question here whether this is a satisfying analysis (see Poschmann 2009; Reis 2012 for revisions). I only want to point out that, if Reis is right, there is the possibility that the respective implicature could be intentionally false. If, on the other hand, the force of the question is triggered by the rising intonation and/or focus on the wh-element, it can be argued that these devices may be deceptively used. Let us now have a look at imperative sentences and optative sentences. (18)

Gib mir den Schlüssel! Give me the key! ‘Give me the key!’

(19)

a.

Hätte ich doch nur den Had SUBJ.I MP MP the ‘Had I only signed the lease!’

Mietvertrag lease

unterschrieben! signed!

54   

   Approaching a linguistic concept of lying

b.

Wenn ich doch nur den Mietvertrag If I MP MP the lease ‘If only I had signed the lease!’

unterschrieben hätte! signed had

Roughly, imperative sentences express the wish of the speaker that a certain future state will happen. For example, in (18) there could be a future state in which the speaker has the keys. Deception is possible with imperatives: The speaker does not have the wish they express, or the keys mentioned do not exist. But lying by using imperative sentences, in fact, appears not to be possible. It is similar with optatives like in the utterances under (19) in which the wish that something were the case is expressed (cf. Grosz 2012). Here the speaker may deceive the hearer in that they do not hold this wish, or in that the lease does not exist in the first place. Again, we would not want to say that the speaker is lying (but see 4.6). Let us now consider exclamative sentences. The crucial question is, whether someone who utters an exclamation like Hat die vielleicht Beine! (‘Boy, the legs she has!’) while not believing that the person referred to has remarkable legs is lying. Interestingly enough, German displays a certain variety of exclamative sentences: (20)

a.

Klingelt Rings

DER aber he MP

heftig! (V1, exclamative accent) vigorously!

b.

DER He

klingelt rings

aber MP

heftig! (V2, exclamative accent) vigorously!

c.

Wie How

klingelt rings

der! he!

(wh-V2)

d.

Wie How

der he

immer klingelt! (wh-VL) always rings!

e.

Dass That

du you

aber auch MP MP

immer so klingeln musst! (dass-VL) always so ring INF must!

For Altmann (1993), the existence of a special accent, called exclamative accent, is one major argument to assume a separate sentence type, the exclamative sentence. This accent, he argues, is neither a normal primary accent nor a contrastive one. Other scholars object, however, that a characteristic type of accent should not be decisive, for the simple reason that accentuation and intonation are not formal (syntactic) devices, but pragmatic ones. These scholars tend to reduce so-called exclamative sentences to either declarative sentences (20a, b) or interrogative sentences (20c, d). I will not go into this debate here (cf. Roguska 2007a,

Interface approach: Lying and sentence types   

   55

Brandner 2010, d’Avis 2013). What is of interest to us is the so-called factivity of exclamative sentences (cf. Abels 2010). On the one hand, factivity is a notion used in connection with presuppositions of certain verbs. Thus, factive verbs such as to realize presuppose their complement: (21)

a. >>

Ken realized that Markus owns an old Mercedes. ‘Markus owns an old Mercedes.’

b. >>

Ken did not realize that Markus owns an old Mercedes. ‘Markus owns an old Mercedes.’

The positive as well as the negative sentence are consistent with the fact, hence presuppose that Markus owns an old Mercedes. Only a special class of verbs shows that behavior, as becomes clear when you compare a verb like to assert. According to Grimshaw (1979), Zanuttini and Portner (2003), and Abels (2010), exclamatives are factive. Thus, exclamatives cannot be answered (cf. 22), they cannot be used as answers (cf. 23), and they are only embeddable by factive predicates (cf. 24): (22)

Speaker A: Speaker B:

Didn’t he eat everything! #Yes, he ate everything.

(23)

Speaker A: Speaker B:

How tall is John? #How tall he is!

(24)

a.

It’s amazing what a fool Bill is.

b.

*It’s possible what a fool Bill is.

While these data is certainly evidence for certain presuppositional effects, it is obvious that exclamatives do not survive the negation test, for the simple reason that the negation test is not applicable. Roguska (2007a) argues that exclamatives are nevertheless presuppositional. But according to her, these presuppositions are not semantic in nature (as one could argue if they passed the negation test), but pragmatic. What is a pragmatic presupposition? It is the assumption that a certain piece of information is already universally known in the situation of an utterance. In the parlance of Stalnaker (2002), every communication has a common ground (CG), i.e., a set of beliefs which are shared by speakers and addressees.

56   

   Approaching a linguistic concept of lying

Moreover, Roguska (2007a: 90) points out that exclamatives do not convey new information; instead, the information (also called ‘the answer’ in her approach) that is conveyed by someone who utters an exclamation is already part of the common ground. What an exclamation contributes to the common ground is only the information that the speaker is surprised with respect to a certain fact. Roguska admits that there are utterances such as Welchen blöden Kerl sie geheiratet hat! (‘What a stupid guy she married!’) whose content may be unknown to the hearer. But in these cases, she argues, the new information is “accommodated” (see Beaver and Zeevat 2007). Accommodation is a technical term first proposed by Lewis (1979: 340): (25)

Accommodation If at time t something is said that requires presupposition P to be acceptable, and if P is not presupposed just before t, then – ceteris paribus and within certain limits – presupposition P comes into existence at t.

In her example, she points out that it is not the ‘answer’ that is accommodated but a weaker proposition, namely ‘Maria hat jemanden geheiratet’ (‘Maria married someone’) (Roguska 2007a: 92). In addition, it is accommodated that the speaker knows the instantiation of the wh-variable. Thus, the assumption that the main point of exclamatives is not to give information but to convey the speaker’s attitude towards some fact, is borne out. One wonders, however, whether the accommodated presupposition is not an assertion. If so, it would make sense to argue that you can lie while using exclamations. Roguska (2007a: 98–103), in her discussion of the assertional approach, concedes that this is a theory-dependent question (her approach fits better into the framework of Stalnaker’s theory of the common ground), and she concedes that an assertional approach may be more suggestive with respect to V1- and V2-exclamatives. I think that the assertional approach is still attractive. Here are three reasons: First, the content of assertions, too, may be known or unknown in the context. Thus, assertions may introduce information unknown to the hearer, or they are used to insist on the validity of a content disbelieved by the hearer. Second, the content of exclamations cannot be cancelled, but it can be denied, just like in the case of assertions: (26)

a.

Hat der aber Beine! #Ich glaube nicht, dass er besondere Beine hat. ‘The legs he has! #I do not believe that his legs are special.’

Interface approach: Lying and sentence types   

b.

(27)

Was der für Beine hat! #Ich glaube nicht, dass er besondere Beine hat! ‘What legs he has! #I do not believe that his legs are special.’

Speaker A: Speaker B:

(28)

   57

Speaker A: Speaker B:

Hat der aber Beine! ‘The legs he has!’ Wieso? Der HAT doch keine besonderen Beine! ‘Why? There is nothing special about his legs.’ Was der für Beine hat! What legs he has! Wieso? Der HAT doch keine besonderen Beine! ‘Why? There is nothing special about his legs!’

Third, reconsider examples like (23), here repeated for convenience, and compare them, as Roguska (2007a: 103) does, with (29): (23)

Speaker A: Speaker B:

How tall is John? #How tall he is!

(29)

Speaker A: Speaker B:

How tall is John? John is extremely tall.

That (23) is not acceptable while (29) is, can – according to Roguska – easily be explained if it is assumed that exclamatives presuppose their contents. But the answer in (23) is simply inadequate because there is no instantiation of the question word. Nevertheless, the exclamative appears to have an assertive content, namely that John is very tall. Abels (2010) also supports the view that the factive meaning of exclamatives is a presupposition. He concentrates on a subclass of exclamatives, namely the what-a exclamative as in (30a) and the how-very exclamative as in (30b): (30)

a.

What a great doctor Ryan is!

b.

How very nice Ryan is!

Abels (2010) defends the classical assumption of Grimshaw (1979), Zanuttini and Portner (2003), and others that exclamatives come along with a factive presupposition that is lacking regular wh-questions. This line of argument includes a rebuttal of recent proposals by Potts (2007), according to which exclamatives

58   

   Approaching a linguistic concept of lying

carry conventional implicatures. Abels (2010: 150) summarizes that (a) “the argument based on the generalization that antifactives do not embed what-a exclamatives and how-very exclamatives” fails; (b) “the argument from the fact that what-a exclamatives and how-very exclamatives cannot be used as answers” is questionable because this might not be a valid criterion for presuppositionaliy”; (c) the test with Really?-replies and Hey, wait a minute!-replies, intended to show that what-a exclamatives and how-very exclamatives cannot be answered, is also implausible because the tests are not watertight. If the information rendered by exclamatives has the status of a presupposition (besides the expressive information that the speaker is surprised, etc.), this presuppositional information nevertheless could be active in lying. This view may be supported by the recent analysis of Zhang (2008) who criticized the relevance-theoretic approach of Simons (2005). Zhang’s main argument is that it is not so much cooperativity and relevance which matter in the computation of presuppositions, but truthfulness. Hence, Zhang (2008: 376) develops certain principles that guide speakers and hearers in the computation of presuppositions. For instance: (31)

a.

Speaker-oriented principle: Presuppose only propositions that you have knowledge about, for truthfulness’s concern.

b.

Hearer-oriented principle: Accept whatever is presupposed by the speaker, for truthfulness’s concern, except when your knowledge forcefully disproves it.

In summary, then, I would like to argue that exclamatives are good candidates for lying besides declarative sentences – both have a certain content that is treated as a fact by the speaker. Whether exclamatives are presuppositional, giving rise to a factive presupposition, or assertional in nature, is still open to discussion. If assertional, we can easily include exclamatives in our view that it is possible to lie while using them. The same applies if they turn out to be presuppositional as the majority opinion suggests: Presuppositions may be false or true, so there exists the possibility to lie to others while falsely presupposing something. In addition, if the speaker does not have the special exclamative attitude they express, they are also deceiving the hearer. This line of argument will be elaborated on in sections 3.8 and 4.6. More generally, we have seen that specific sentence types may be associated with meaning effects that could have an effect of their lying potential. Thus, V1-exclamatives are associated with narrative uses which seem to exclude prototypical assertions. Wh-interrogatives and exclamatives are associated with impli-

The case of relative clauses   

   59

catures/presuppositions that could be conveyed in an untruthful manner. While it seems reasonable to view the declarative sentence as the prototypical formal device used for lying, it appears that other sentence types also contain certain relations to lying.

2.5 The case of relative clauses The interface approach suggested in the preceding section demands a certain view on speech-act assignment, i.e., the assignment of illocutions to grammatical devices such as the sentence type. The question of lying with regard to sentence types gets still more complex when embedded sentences such as the relative clause are considered. The following examples illustrate the prominent distinction between restrictive and appositive relative clauses: (32)

a.

Das Jurymitglied, das in seinem Urteil schwankte, wollte Kaffee. The jury member that in his judgment shifted wanted coffee. ‘The jury member who shifted his judgment wanted coffee.’

b.

Das Jurymitglied, das ja in seinem Urteil schwankte, wollte Kaffee. The jury member that MP in his judgment shifted wanted coffee. ‘The jury member who shifted his judgment wanted coffee.’

Both utterances may be used by a liar. Utterance (32a) is a lie if the speaker does not believe that the specific jury member who shifted his judgment wanted coffee (restricted reading) but still asserted this (with the intention to deceive the addressee). The noun phrase the jury member who shifted his judgment picks out a particular member of the jury. In (32b), the relative clause may be used to lie in a double way. First, if the jury member did not want coffee while this is what the liar falsely asserted, this would count as a lie. Second, if it was not the case that the coffee-demanding jury member was shifting his judgment, and the liar knew it, this is also a lie. What is it that makes appositive relative clauses useful lying instruments? Their pragmatic properties have been repeatedly mentioned in the literature, but there is no systematic pragmatic theory with interfaces to phonology, syntax, and semantics available (cf. Blühdorn 2008, Holler 2007, Birkner 2008: 13–100). Several putative properties of appositive relative clauses have been pointed out, for instance that they are in the background of the sentence containing them, that they deliver subsidiary information (‘Nebeninformation’), that their content is presupposed, or that they introduce a further meaning dimension (Potts 2005).

60   

   Approaching a linguistic concept of lying

It has also been assumed that appositive relative clauses possess a separate focus-background structure (Holler 2007: 253ff.), a separate proposition, and an autonomous illocution (Holler 2007: 254f.). While this is a quite puzzling array of concepts, the idea that appositive relative clauses have their own illocutions and/or that they convey certain presuppositions fits to my overall assumption that it is possible to lie by using them. As evidence for an appositive relative clause, the following criteria have been proposed: (a) the kind of NP to which the appositive relative clause is related to, viz. personal pronoun 1. and 2. Ps., proper names, demonstratives with dies- und jen-, cf. Birkner (2008: 34f.); (b) the selection of certain lexical indicators such as the modal particle ja, the discourse particle übrigens (‘besides’), and sentence adverbs like bekanntlich (‘as everyone knows’); their insertability is often used as an appositivity test; (c) the referential function of the NP to which the appositive relative clause is related; in restricted relative clauses, the set of possible referential object is restricted, while appositive relative clauses specify the already identified referent of the NP; (d) the scope of the determiner of the NP; restrictive relative clauses are in the scope of the NP’s determiner, while appositive relative clauses are not; (e) the referentiality of the relative pronoun (in restrictive relative clauses, it is a bound variable, in appositive relative clauses, it is referential) (Blühdorn 2008, Holler 2007). While these elements certainly influence the reading of appositive relative clauses, they are not sufficient to determine the intended reading. For instance, the ambiguity of Sie hat vier Töchter, die Rechtsanwältinnen wurden (‘She has four daughters who became lawyers’) needs a specific context of utterance for resolution (Birkner 2008: 33). Moreover, as Loock (2007) has shown, we have to distinguish at least three different pragmatic functions, namely what he called (i) continuative appositive relative clause, (ii) relevance appositive relative clause, and (iii) subjectivity appositive relative clause. And these types of appositive relative clauses are connected with certain conversational implicatures in turn. As already indicated, the most interesting question related to lying is whether appositive relative clauses are illocutions and/or presuppositions  – both pragmatic entities suited for lying. A third possibility (which I do not find plausible as I will point out) is that appositive relative clauses are conventional implicatures sensu Potts (2005). My main argument draws on the meaning and function of the German modal particle ja which appears to indicate appositivity and is hence used as a litmus test for appositive relative clauses. This particle has a twofold nature: On the one hand, it refers to information already known in the context of an utterance. On the other hand, it is possible to introduce new information with ja. There are numerous descriptions of ja in the German literature. Most scholars assume that

The case of relative clauses   

   61

ja signals that a proposition is already known to the hearer or should be taken for granted by them (for a review of the literature, see Hoffmann 2008). I follow the definition of Lindner (1991) rendered in (33) (cf. Meibauer 1994): (33)

(It is necessary that) If the speaker uses ja in an illocution type IT referring to a proposition p, then s/he assumes at the time of speaking t that it is not the case that there is a proposition q in the set of propositions activated at t such that p is not true. (Lindner 1991: 178)

Now I would like to argue that ja is an element that either signals a presupposition or indicates the need for the accommodation of a proposition. In the latter case, ja may also be taken as an illocutionary indicator for an assertion. (This explains why appositive relative clauses have been associated with presuppositions as well as with assertions.) Both uses are compatible with their reading as appositive relative clauses. This double nature finds further support when one considers the argument presented by Freienstein (2008) with regard to the sentence adverb bekanntlich (‘as is known’) that is also used as an appositivity indicator. It seems that appositive utterances violate the second maxim of Quantity (Grice 1989a), requiring that a contribution should not be made more informative than necessary. However, the speaker is often unsure about what may be presupposed as already known. Hence, as Freienstein (2008: 148) explains, if the speaker does not use the appositive relative clause, he may be underinformative, but if he uses it, he may be overinformative. Also, the speaker may implicate that his audience should be acquainted with a certain content or that the audience should take a certain content for granted. Another strategy in using appositives is to implicate that a certain content should be considered to be shared knowledge in order not to be held responsible for that content, especially in the case of blaming someone (… Ken who is well-known as a bad researcher …+> ‘It is shared knowledge that Ken is a bad researcher’). It is goes without saying that this is a very good strategy for a potential liar (who secretly believes that Ken is excellent) because he leaves it to the hearer to be duped by way of implicature. We will come back to this case when we talk about the case of falsely implicating (section 4.5). As for the presuppositional analysis, I side with Schlenker (2007) who develops a presuppositional analysis for expressive devices such as ethnophaulisms (‘ethnic slur terms’) like engl. honky and T/V-pronouns like French tu. These presuppositions are (a) indexical, i.e., they relate to a specific context, (b) attitudinal, i.e., the presupposition conveys a specific attitude of the speaker, and (c) shiftable, i.e., the context of evaluation may differ from the actual context. All of these properties are also valid for modal particles. Thus, ja may refer to past

62   

   Approaching a linguistic concept of lying

or actual contexts, expresses an attitude of the speaker (namely that something should be considered as uncontroversial), and is shiftable, cf. Ken behauptet (ja), dass Barbie (ja) Peter geküsst hat (‘Ken asserts (MP) that Barbie (ja) kissed Peter’). Schlenker (2007: 242) observes that there may be presuppositions that introduce new information, for example, if someone utters I met a honky the presupposition ‘speaker believes that white people are despicable’ becomes common ground. Similarly to the utterance of Ken ist ja jetzt verheiratet (‘Ken is MP married now’), someone may inform somebody about the fact that Ken recently got married (see the case of informative exclamations above). Because of this twofold nature of the modal particle ja, I do not believe that appositive relative clauses are conventional implicatures in the sense of Potts (2005). Conventional implicatures in the sense of Potts (2005) are conceived of as being logically and compositionally independent of what is being said; they are, in the parlance of Potts (2005), ‘at-issue’-implications (cf. Amaral, Roberts, and Smith 2007, McCready 2010). But modal particles such as ja do not signal implications but attitudes of the speaker towards a proposition (only in this sense they may be regarded as “expressive”). This proposition may be either presupposed or accommodated. In the latter case, the information appears as asserted. If presupposed, the information goes together with the assumption of its truthfulness.

2.6 Conclusions What we have found is evidence that there are several sentence types that encode information that may be used for lying. Besides the declarative sentence that entertains a natural relationship to assertions, as we will see in more detail in the next chapter, it is the exclamative sentence type that is of interest here. Exclamatives, so I have argued, are connected with either presuppositions or assertions, because they contain a regular factive element. If this factive element is a presupposition and the presupposition is intentionally presented as true while it is in fact false, we have a case of lying. (Note that I will argue that intentionally false presuppositions and implicatures should be included into a general definition of lying, see chapter 4.) We faced a similar case with appositive relative clauses, which also have been argued to display presuppositions or illocutions. Focusing on the appositivity marker ja, I showed that the meaning of ja is compatible with either analyses. Since our main concern is to describe lying at the semantics-pragmatics interface, we finally have to explain, how the morpho-syntactic component interacts with the minimalist model we discussed at the beginning of this chapter. Most semantics-pragmatics interface models I know of (e.g., Levinson 2000, Recanati

Conclusions   

   63

2010, Garrett and Harnish 2007) neglect this connection of semantics and pragmatics to grammar or simply take it for granted. See Table 3: Table 3: Presemantic and postsemantic pragmatics (Levinson 2000: 188)

Compositional Semantics

Indexical Pragmatics

Output: Semantic Representations

Gricean Pragmatics 1 disambiguation, fixing reference, generality-narrowing, etc.

Semantic Interpretation model-theoretic interpretation

Output: sentence-meaning, proposition expressed

Gricean Pragmatics 2 indirection, irony and tropes, etc.

Output: speaker-meaning, proposition meant by the speaker

64   

   Approaching a linguistic concept of lying

Hence, the standard minimalist picture presented by Recanati (2010) (see Table  2, p.  38) has to be fleshed out: Before the level of sentence meaning is reached, the morpho-syntactic machinery has done its work. Lexical items have been drawn from the lexicon and have been inserted into syntactic structures. Furthermore, sentences with their compositional semantics get a phonological structure, and this phonological structure is interacting with syntax as well as with semantics. Even in the Neogricean approach of Levinson (2000: 188), the lexicon, morpho-syntax and phonology are not represented. Furthermore, we need an illocutionary apparatus that is able to connect different sentence types, together with their specific semantic modality, to appropriate illocutions. Such a connection is also not evident from the above model. And last but not least, all acts of production and comprehension are related to specific situations of an utterance, so the context must be taken into account. This in turn leads to the question of how the whole model is related to the context (see Meibauer 2012b). With these considerations in mind, we now turn to the analysis of assertions.

3 Assertion and lying 3.1 Introduction In this chapter, we will discuss several aspects of a proper definition of assertion. Recent literature shows that there are many attractive proposals beside the classic definition of Searle (1969). We cannot, however, discuss all these aspects in depth; even the recent monograph on assertion by Jary (2010) is not comprehensive. We have a more modest aim, namely to propose a definition of assertion that fits into standard models of semantics-pragmatics interactions (such as the standard minimalist models discussed by Levinson 2000 or Recanati 2010), and that is useful for an intelligible definition of lying.

3.2 Assertion and statement Most people think of lying as essentially being a deceptive or insincere assertion. But what is an assertion? “Assertion” is one of the most basic notions in the philosophy of language, yet there are quite different theories about assertions on the market as Pagin (2007) makes clear in his excellent contribution on assertion in the Stanford Encyclopedia of Philosophy. In the present section, we will have a closer look at the notion of assertion from the point of view of speech-act theory, i.e., we will focus on the  – as Jary (2010) calls them  – “speech-act traditionalists” Searle (1969), Bach and Harnish (1979), and Alston (2000). These speech-act traditionalists assume that assertion is but one out of a number of speech act types, on a par with directives, commissives, expressives, etc. So-called “speechact fundamentalists”, on the other hand, like Brandom (1994) and Barker (2004), ascribe a more fundamental position to assertions, e.g., that assertion has also an impact on other types of speech acts. While these latter approaches may also have their merits for an analysis of lying, I think that there are good reasons to stick to the speech-act traditionalists because their approaches are more useful from a linguistic point of view. In short, speech-act traditionalists accept the fact that speech acts are systematically related to the forms of a language, while speechact fundamentalists focus on philosophical or psychological aspects of assertive meaning (see Jary 2010: 16–30 for a review). In the philosophy of language, the constant source of linguistic concepts of speech acts, there are four basic approaches to assertion (Brown and Cappelen 2011: 2–4):

66   

   Assertion and lying

(a) “Assertions are those sayings that are governed by certain norms – the norms of assertion.” (b) “Assertions are those sayings that have certain causes.” (c) “Assertions are those sayings that have certain effects.” (d) “Assertions are those sayings that are accompanied by certain commitments.” There is a fifth view added – the “debunking” view of Cappelen – who insists that “there is no one correct way to characterize the extension of assertion” (Brown and Cappelen 2011: 4). Cappelen may have a point here; however, in order to be able to describe lying, we certainly need a robust concept of assertion, even if it is only correct with respect to this task. The first approach mentioned above is associated with the founders of speech-act theory, John L. Austin (1962) and John R. Searle (1969). Their basic idea is to define illocutions with the help of felicity conditions. Thus, a speech act of assertion is only felicitous, i.e., fulfils a certain norm, if its propositional content is true, if the speaker believes that it is true, etc. (see Turri 2013 for experiments concerning the norm of factivity). The second view is ascribed, for instance, to Bach and Harnish (1979), who characterize speech acts in terms of expressing a certain attitude; that is, a certain speaker attitude, i.e., belief, is the cause for the speech-act category of assertion. The third approach, as assumed by Stalnaker (1979, 2002), is to define assertions by the effects they have on the ongoing communication. Finally, the fourth approach (e.g., by Brandom 1994) characterizes assertions in terms of the commitments they are associated with, for example the commitment to present evidence for the truth of the asserted content. It appears that all of these approaches put their fingers on an important aspect of assertion. Pagin (2007: 1) starts with a simple definition: “An assertion is a speech act in which something is claimed to hold.” For example, in the speech acts in (1)–(2), the contents of these assertions may be fixed as under (3)–(4): (1)

There is a traffic congestion on Adenauer Bridge.

(2)

Markus has tooth ache.

(3)

‘that there is a traffic congestion on Adenauer Bridge at t’

(4)

‘that Markus has tooth ache at t’

Assertion and statement   

   67

These contents are usually called ‘propositions’. Propositions are typically described by that-sentences. Since propositions, as we will see later in more detail, may be evaluated with respect to their truth or falsity, we need a time variable t to be instantiated in the context of utterance. Searle (1969) proposed the following definition of assertion. (5)

Assertion: Rules for the use of illocutionary force indicating devices (Searle 1969)

Propositional content rule What is to be expressed is any proposition p. First preparatory rule S has evidence (reasons etc.) for the truth of p. Second preparatory rule It is not obvious to both S and H that H knows (does not need to be reminded of, etc.) p. Sincerity rule S believes p. Constitutive rule Counts as an undertaking to the effect that p represents an actual state of affairs. Let us shortly comment on these rules. In Searle’s speech-act theory, there is a strong link between verbal elements with a specific meaning and certain types of speech acts. Thus, the declarative sentence, as in (1) and (2), or, more at the heart of speech-act theory, the performative verb to assert are devices that can signify an assertion. The propositional content rule, then, says that every proposition p can be asserted. There is no restriction in that respect, for example, that a proposition must relate to a future state of affairs (as with a promise) or past state of affairs (as with an oath). In addition, the rule states that assertions must have a certain propositional content. Simply uttering I assert without any specification about what is being asserted does not count as a successful assertion. The first preparatory rule tells us that you must not assert something if you have no evidence for the truth of p. Hence, uttering There is a traffic congestion on Adenauer Bridge, but I have no evidence for that state of affairs will not do. And the obvious, so goes the second preparatory rule, should not be asserted. For example, while we are stuck in the traffic congestion on Adenauer Bridge, it makes no sense to utter (1). (Or, if it is just the obvious that is being told, this does not count as an assertion.) The sincerity rule guarantees that in the case of an assertion, the proposition must be believed by the speaker. Therefore, it is not felicitous to utter There is a traffic congestion on Adenauer Bridge but I don’t believe this. The constitutive rule comprises the essence of this type of speech act: the point of assertions is to represent something. Why is this rule called “constitu-

68   

   Assertion and lying

tive”? The idea is that, “without this rule the practice of assertion would not exist”, and “the existence of the undertaking is an institutional fact created by the utterance”, as Pagin (2007: 16) puts it. (In contrast, regulative rules describe the state of affairs that exist independently of the rules.) Influential as it was, several critical remarks may be made. These are related to the overall architecture of Searle’s approach with its distinction of different types of rules. Among other things, Harnish (1990: 173–177) presented three arguments. First, it is not always clear why a particular speech-act property is subsumed under a particular rule. For instance, in the case of promises, the speaker’s assumption that a future action by the speaker is in the interest of the hearer belongs to the preparatory rules. Why not put this into the constitutive (essential) rule where it would be useful in the discrimination of threats? Second, it is unclear which rules are necessary for the constitution of a speech act. There are speech acts without a propositional content or a sincerity requirement, e.g., greetings. Therefore, it seems that preparatory rules and constitutive rules are the most important ones. Third, one should make a distinction between those rules which are absolutely necessary for the felicity of a speech act, and those which are not. For instance, the preparatory rule for promises demands that the speaker thinks that a future action of the speaker is in the interest of the hearer. So, when this assumption is being mistaken, and the hearer does not accept the promise, the speech act would be felicitous but nevertheless somehow defective. Jary (2010: 9) summarizes that “on Searle’s account, an assertion that P will be unhappy if it is made under circumstances in which it is obvious to both the speaker and the hearer that the hearer already knows that P, but it will still be an assertion, for it will still count as an undertaking that P is true.” In such a case, the utterance violates the second preparatory rule, so I do not think that this would count as an assertion. If an “undertaking” in the sense of the constitutive rule is made while knowing that the second preparatory rule is not instantiated, I would suspect that the speech act is not an assertion but a statement (both speech acts belonging to the class of assertive speech acts). The definition of assertives proffered by Bach and Harnish (1979) is more easily embeddable in a Gricean theory of communication: (6)

Assertives (Bach and Harnish 1979) In uttering e, S asserts that P if S expresses: i. the belief that P, and ii.

the intention that H believe that P.

Assertion and statement   

   69

In contrast to Searle’s analysis, Bach and Harnish draw attention to the fact that by asserting, we do not only want to represent a state of affairs, but, in addition, we want the addressee to share our beliefs with respect to this state of affairs. This is indeed an aspect of assertion that is lacking in Searle’s analysis. As for “expressing a belief”, a further comment is in order. For Bach and Harnish (1979: 15), the definition in (7) holds; “R-intend” is short for ‘reflexively intend’: (7)

Expressing an attitude For S to express an attitude is for S to R-intend the hearer to take S’s utterance as reason to think that S has that attitude.

Thus, speakers communicate hearer-directed intentions concerning their beliefs, and the recognition of the speaker’s intention by the hearer is crucial. Pagin (2007: 13) argues that, e.g., during an interrogation, “a speaker can clearly make an assertion even if the addressee has a prior conviction that the speaker is dishonest” and “she will not then intend the interrogators to take her utterance as evidence for belief.” I disagree. It is not possible to utter something like There was a traffic congestion on Adenauer bridge but (as I know that you do not trust me) I do not want you to believe that I believe that this was the case. In the same vein, Jary (2010: 10) diagnoses a “problem faced by Bach and Harnish’s definition: it does not class as assertions those cases where the speaker has no intention of having his audience believe the content of his statement.” For instance, a pupil answers Paris is the capital of France to the teacher’s question What is the capital of France? Here, the pupil asserts something, but she has no intention to cause the respective belief in the teacher. Again, I think that this would be a case of a statement. Jary (2010: 10–11) argues that there is a dilemma here: on the one hand, the informativity of assertions should be captured; on the other hand, the possibility of non-informative assertions should be acknowledged. His idea to make a distinction between ordinary and unordinary assertions is not appealing since it looks like a terminological trick. What should be distinguished are situations in which information is being asserted (as new information for the hearer) or just stated (something that is already known is made manifest). I will come back to this point shortly. Alston (2000: 120), still in the tradition of speech-act traditionalists, proposes the following analysis schema for assertive acts:

70   

   Assertion and lying

(8)

Analysis schema for assertions (Alston 2000) U asserted that p in uttering S iff: 1. U R’d that p. 2.

S explicitly presents the proposition that p, or S is uttered as elliptical for a sentence that explicitly presents the proposition that p.

In this schema, “U” abbreviates “speaker (utterer)”, “p” is a variable for propositions, “S” stands for “sentence”, and “U R’s that p” means “U takes responsibility for its being the case that p”. What is new and must be explained is the idea that in asserting, the speaker takes “responsibility for (its being the case that) p” (Alston 2000: 55). Alston (2000: 54) gives the following explanation, which makes his normative approach crystal clear: “’Take responsibility’ must be understood in a special way. The idea is not that U (utterer, speaker) took responsibility for state of affairs C in the sense that he was prepared to acknowledge that he brought C into existence. It is, rather, like the way in which, when I become head of department or agency, I take responsibility for the efficient and orderly conduct of its affairs, including the work of my subordinates. I am responsible for all that work, not in the sense that I have done it all myself, but in the sense that I am rightly held to blame if the work is not done properly. I am the one who must “respond” to complaints about that work.”

More formally, the idea is expressed in (9): (9)

U took responsibility for (its being the case that) p (Alston 2000: 55) In uttering S, U R’d that p – In uttering S, U knowingly took on a liability to (laid herself open to) blame (censure, reproach, being taken to task, being called to account), in case of not-p.

While there is a certain affinity to Searle’s first preparatory condition (“S has evidence (reasons etc.) for the truth of p.”) or Grice’s second sub-maxim of the maxim of Quality (“Do not say that for which you lack adequate evidence.”) to be noticed, the accent is here on the social and normative aspects of acting. That is, the social consequences of violating the responsibility obligation is built into the definition of the speech act. Note that with respect to our enterprise to define lying, this move is quite welcome for it explains to a certain extent harsh moral reactions against the liar. Another move in Alston’s analysis is that he explicitly includes the possibility of asserting something by ellipsis. He comments (Alston 2000: 120):

The sincerity condition   

   71

“[…] we must recognize the possibility of asserting that p without uttering a sentence that itself explicitly presents the proposition that p. Surely when I assert that I had soup for lunch by saying ‘soup’, in answer to the question ‘What did you have for lunch today?’, my sentence does not explicitly present the proposition that I had soup for lunch today. Hence we must allow for derivative cases of assertion in which what one utters is used elliptically for a sentence that does explicitly present the proposition in question.”

Whether ellipsis should be built into a definition of assertion, is unclear to me, because we need a linguistic theory of ellipsis anyhow. But the effect of being able to say that someone asserts (and hence might lie) while using elliptical constructions is of course very welcome. Let us now turn to the distinction between assertions (Germ. ‘Behauptungen’) and statements (Germ. ‘Feststellungen’). The latter are usually subsumed under the label of assertives where assertive is a kind of class label. For example, Bach and Harnish (1979: 42) list verbs denoting illocutionary acts like affirm, allege, assert, aver, avow, claim, declare, deny (assert … not), indicate, maintain, propound, say, state, submit under the label of “Assertives (simple)”, these being a part of the super class of Constatives in turn. What is, then, the difference between asserting and stating  – a difference that intuitively appears to be relevant for our attempt at defining lying? Grewendorf (1982: 135) proposes that for assertions, evidence for their truth must be given only in the future, whereas with statements, one can rely on the assumption that they are the result of some prior investigation of the facts. Assertions have a provocative character, while statements have a resultant character. Thus, it is possible to explain with respect to a given piece of information: “This utterance of mine was not meant as an assertion, only as a statement.” Similarly, in the context of the didactic question What is the capital of France?, when the pupil gives the answer the teacher already knows, they simply state that Paris is the capital of France. In sum, then, while focusing on different aspects of asserting, all of these approaches have a convincing aspect: When asserting, the speaker expresses, in a responsible manner, their belief that p, and wants the hearer to recognize and take over this belief that p.

3.3 The sincerity condition Truth and truthfulness constitute an epistemic relation  – what the speaker believes and knows  – whereas truth and sincerity constitute a pragmatic relation – what the speaker truthfully utters. In Searle’s speech-act theory, there is a

72   

   Assertion and lying

separate rule of sincerity that requires, as we have already seen, that the speaker believes p in the case of assertion. Lying, then, is for him an act of insincere assertion. Let us have a look at the advantages and shortcomings of this approach. The background of Searle’s analysis is the famous analysis of promising in the third chapter of his “Speech Acts” (1969). There, he proposes a sincerity condition for promises. This sincerity condition is in turn related to the sincerity rule that describes the correct use of the illocutionary force indicating device: Pr is to be uttered only if S intends to do A. See Searle (1969: 63) and (5) above. (10)

Sincerity condition for promises (Searle 1969: 60) S intends to do A.

Now it is well-known that sincerity, i.e., the actual intention of the speaker to do A in the future, is not actually existent in many acts of promising. Barbie promises to come to the party, but she does not intend to come because she already has a date with Ken. Why, then, does the utterance I promise to come to the party nevertheless count as a promise? Searle (1969: 62) proposes that “[…] to allow for insincere promises, we need only to revise our conditions to state that the speaker takes responsibility for having the intention rather than stating that he actually has it.” Thus, he comes up with a revised sincerity condition: (11)

Revised sincerity condition for promises (Searle 1969) S intends that the utterance of T will make him responsible for intending to do A.

However, several criticisms have been put forward against this amendment. I will sketch two of them. The first is due to Falkenberg (1990: 134) who observes that the insertion of (11) into the original analysis schema for promises leads to the question what the relation between the sincerity and the essential rule is. And in fact, it is not easy to see how the sincerity condition and the essential condition in (12) are distinct: (12)

Essential condition for promises (Searle 1969: 60) S intends that the utterance of T will place him under an obligation to do A.

The difference between responsibility and obligation is not explained by Searle; however, since his aim is to provide an analysis that “is neutral on the question whether the promise was sincere or insincere” (Searle 1969: 62), it may be suspected that an extra condition like (11) is superfluous. Insincerity, then, may be something that cannot be accounted for by some condition or rule. An insincere

The sincerity condition   

   73

promise may be conceived as an act in which the speaker deceives the hearer about their true intention. The second problem has to do with Searle’s distinction between regulative and constitutive rules. According to Searle, constitutive rules are rules that constitute or bring about certain behavior, while regulative rules regulate a behavior that exists independently of the rule. Constitutive rules are, for example, the rules of football or chess “that create the very possibility of playing such games”. Regulative rules, for example, are “rules of etiquette”, regulating “inter-personal relationships which exist independently of the rules” (Searle 1969: 33). Searle takes the essential condition as a constitutive rule, while for him the sincerity condition is a regulative rule. But if the revised sincerity condition has a content that is at least very similar to the content of the essential condition, the distinction between regulative and constitutive rules becomes blurred, too. Now let us turn to the Searlean analysis of lies. Searle views lies as insincere assertions, but in this case, there is no revised sincerity condition given. The basic problem is, of course, whether it makes sense to define lies as assertions (albeit insincere ones) in the first place. Thus, Dietz (2002: 84) (rhetorically) asks: “Wie kann man mit einer unaufrichtigen Behauptung überhaupt noch einen regulativen Regelverstoß im Sinne des Lügens begehen, wenn jede Behauptung ohnehin als Ausdruck des psychischen Zustands ‘glauben, daß p’ gilt?” ‘How it is possible at all to violate the regulative rules when lying if every assertion is an expression of the psychological state ‘believe that p’ anyway?’ [My translation, JM]

Dietz’ solution is to understand lies as an “eigenständiges Handlungsmuster” (p. 87) (‘autonomous pattern of action’). While lying may be a “pattern of action”, it is not clear to what extent it is “autonomous”. Indeed, lying essentially has to do with asserting the truth of a propositional content. In order to relate lying to asserting and to circumvent the contradiction with the sincerity condition of assertions, a revised sincerity condition similar to the one proposed with respect to promises will do: (13)

Revised sincerity condition for assertions (my proposal) S intends that the utterance of T will make him responsible for the truth of p.

As pointed out above, Searle (1969) gives no additional revised sincerity condition for lies. However, in Searle (1979: 62), in the context of a paper on “The logical status of fictional discourse”, we find a definition of assertion that includes a new version of the sincerity rule:

74   

   Assertion and lying

(14)

Assertion (Searle 1979b: 62) 1. The essential rule: the maker of an assertion commits himself to the truth of the expressed proposition. 2.

The preparatory rules: the speaker must be in a position to provide evidence or reasons for the truth of the expressed proposition.

3.

The expressed proposition must not be obviously true to both the speaker and the hearer in the context of utterance.

4.

The sincerity rule: the speaker commits himself to a belief in the truth of the expressed proposition.

In this definition, a new illocutionary notion is introduced, namely ‘commitment’. Searle argues in this paper that assertions in fictional discourse are assertions indeed, although they do not comply with the above rules; but they are pretended acts, because “an author of fiction pretends to perform illocutionary acts which he is not in fact performing”(p. 66). Dietz (2002: 87) does not speak of “pretended” acts, but of declared (“deklariert”) acts; while fictional discourse contains signals that point to the fictional character, lies do not contain such pointers; they are covered acts. Dietz (2002: 87) worries that according to Searle’s approach it is unclear how a speaker should ever detect a lie, since for him a lie does not follow its own rules, while fiction does. In Dietz’ approach, only when lying is conceived as an autonomous pattern of action, it can be defined as a rational violation of the sincerity rule. I think, however, that this is a misunderstanding of Searle’s approach. If a hearer detects that a speaker asserts p while the speaker does not believe that p, then this is evidence enough. There is no need to assume a special pattern of action leading to a separate type of speech act.

3.4 Perlocution In Austinian speech act theory, a threefold distinction between locution, illocution, and perlocution is made. Thus, Austin (1962: 101) defines: “Saying something will often, or even normally, produce certain consequential effects upon the feelings, thoughts, or actions of an audience, or of the speaker, or of other persons.” For instance, if Ken utters Don’t smoke! to Barbie, this utterance contains the words don’t and smoke (locutionary act), but it is also a command or a warning (illocutionary act), and moreover, if as a consequence of that utterance,

Perlocution   

   75

Barbie does not smoke, the speaker has convinced her not to smoke (perlocutionary act). Perlocutionary acts are thus intended effects of an utterance. This concept has also been applied to lying. Thus, Reboul (1994: 297), similar to Dietz (2002), finds that there is something paradoxical about the definition of lie in speech act theory (where “perlocutionary act” refers to the effect of being lied to): (15)

The paradox of lies in speech act theory (Reboul 1994: 297) “If a speaker produces an utterance which is a lie, it is necessary for the success of the lie that the illocutionary act of assertion should be successful. But if the perlocutionary act of lying is successful, then the illocutionary act of assertion is not successful.”

Her basic idea is to analyze lying as a perlocutionary act. Lies cannot be illocutionary acts, because lies are not conventional, yet conventionality is an important trait of illocutionary acts. Although lies correspond to the illocutionary act of asserting, they themselves are not assertions. Instead, they are perlocutionary acts. Reboul (1994) draws on Cohen (1973) who distinguishes between direct perlocutions (which may but do not necessarily arise) and associated perlocutions (they must arise, if the illocutionary act is felicitous). With respect to this distinction, lies are direct illocutions, since a lie does not automatically follow from an assertion. The intended perlocutionary effect is that the hearer is lied to, i.e., the hearer believes that the speaker believes p, and because of this trust into the speaker’s judgment, the hearer believes p too. Moreover, Reboul (1994: 297–98) proposes to loose the connection between a lie and insincerity, in that every insincere speech act is considered as a lie. The following definition of a lie is a proposal along these lines: (16)

Definition of lie (Reboul 1994: 297) “A lie is any utterance corresponding to the performance of an illocutionary act which has a sincerity condition, where: a) the sincerity rule has not been complied with; b)

the speaker has the intention that his hearer believes that the speaker has respected the sincerity condition.”

The main criticism with regard to the concept of a perlocution is whether one should understand possible effects of an illocutionary act as a separate act performed in addition to the locutionary and the illocutionary act. While it is plau-

76   

   Assertion and lying

sible that some types of acts, e.g. persuasion, aim at changing the hearer’s mind, this can never be predicted. Thus, there is no proof that there is such a thing as a perlocutionary act, only that a certain intended effect belongs to the semantic concept of persuading someone. Therefore, the concept of perlocution has been neglected in speech-act theory (cf. Marcu 2000). The effect of being lied to or not, may be reconstructed in dynamic approaches to conversation, be it in a Stalnakerian framework or in a discourse-related framework. Also the proposal in (16) is problematic since we do not want to speak of an insincere promise or warning as a lie. If such illocutionary acts are conveyed in an insincere way in order to mislead the hearer, we may handle them as deceptive acts.

3.5 The maxim of Quality While it is evident that the sincerity condition in Searle’s theory of speech acts has similar effects as the maxim of Quality in Grice’s theory of implicature, the relation of the two has never been worked out in detail. Since we need both pragmatic theories for a linguistic analysis of lying, we will discuss this relation in the upcoming section. As is well-known, Grice’s (1989a) theory of implicature contains – besides the maxims of Quantity, Relevance, and Manner – a maxim of Quality: (17)

Maxim of Quality (Grice 1989a) Try to make your contribution one that is true. 1. Do not say what you believe to be false. 2.

Do not say that for which you lack adequate evidence.

This maxim looks so simple that it has seldom been analyzed in any detail. Look at the super-maxim “Try to make your contribution one that is true”. This maxim reflects a certain presumption about cooperative behavior. As results of the Cooperative Principle, there is, in every conversation, an “accepted purpose or direction of the talk exchange” in which speakers are engaged. And this purpose requires a certain make-up of conversational contributions. To make contributions that are false, or to mix up false and true contributions, is sort of an irrational behavior. Grice seems to accept that, very often, it is not an easy task to make true contributions – but one should at least “try” to do so. The first sub-maxim requires that speakers should not say what they believe to be false. This sounds like a rule not to lie, but there is no hint at the deceptive intent connected to lies. The second

The maxim of Quality   

   77

sub-maxim requires that speakers – as it may be supposed in the case of assertive acts – should provide “adequate evidence”. This rule echoes Searle’s preparatory condition, while the first sub-maxim bears some relation to the sincerity rule. Two questions arise: First, why does Grice propose the maxim of Quality? Second, what is the relation between the Gricean maxim of Quality and its speech-act theoretic counterpart? A very general reason for assuming a maxim of Quality certainly is that truthfulness has to do with cooperation. This is not to say that lying is always uncooperative, since there are prosocial lies. Yet it is a rational behavior of speakers to rely on the assumption that others speak truthfully. Furthermore, Grice needs the maxim of Quality in order to account for irony, metaphor, and sarcasm – tropes that are analyzed as conversational implicatures – and for belief implicature in assertions: (18)

(19)

a.

He is a fine friend.

+>

‘He is not a fine friend.’

b.

You are the cream in my coffee.

+>

‘You are my best friend.’

There is life on Mars. +> ‘Speaker believes that there is life on Mars.’

According to Levinson (1983), examples (18) are a case of exploitation of the maxim of Quality, while (19) is a case of observation. We will postpone a detailed discussion of irony to section 5.4. Many researchers criticized Grice’s approach since it captures the ambiguity of utterances (18) quite well but is not able to show the exact derivation of the intended meanings, let alone the differences between the tropes. And in the recent approaches of Levinson (2000) and Carston (2002), the maxim of Quality or some related principle plays no role at all. This is not to say that Levinson and Carston delivered the final deathblow to Quality. Quite on the contrary, I have argued elsewhere (Meibauer 2006) that it might have a comeback. The second question mentioned above was what the difference between sincerity and quality is. Let us have a closer look at (19). The implicature should solve Moore’s Paradox, i.e., the observation that an utterance like (20) is weird: (20)

*There is life on Mars but I don’t believe it.

Recall that conversational implicatures should be reconstruable, cancellable, and context-variable. Since cancellability obviously does not work, see (20), we are left with context-variability. The context which matters here is the context of lying. Hence it may be asked whether there exists a thing such as a belief implicature. If the belief that p is built into a definition of an assertion, a separate

78   

   Assertion and lying

belief implicature appears to be superfluous. But the maxim of Quality may be helpful in the enterprise of getting rid of the sincerity condition. When sincerity is equated with the maxim of Quality, then there is no use of expressing it additionally in the conditions for assertion. A recent attack on the maxims of truthfulness has been put forward by Wilson and Sperber (2002). Their overall aim is to show “that language use is not governed by any convention or maxim of truthfulness in what is said.” (Wilson and Sperber 2002: 583) And they hasten to add: “Whatever genuine facts such a convention or maxim was supposed to explain are better explained by assuming that communication is governed by a principle of relevance.” With the term “maxim of truthfulness”, they refer in particular to the first Gricean sub-maxim “Do not say what you believe to be false.” Sperber and Wilson present several arguments against the operation of this maxim. First, as Grice himself pointed out, there are several prima facie categories of violation of the maxim of Quality: (i) lie (covert violation), (ii) jokes/fiction (suspension/opting out), (iii) metaphor, irony (overt violation/flouting). But these may be treated along Gricean lines. However, in a more general perspective, Wilson and Sperber (2002: 586) urge that “many, if not most, of our serious declarative utterances are not strictly and literally true, either because they are figurative, or simply because we express ourselves loosely.” Second, Grice (1989a: 34) argues that in the case of irony and metaphor, the speaker “has said or has made as if to say” something he did not believe. But if they only simulated the saying, there is no need to appeal to a maxim of truthfulness that is basically related to the truth of ‘what is said’. Thus, the fundamental attack of Wilson and Sperber is against the Gricean distinction between what is said and what is implicated. For them, all utterance meanings, and especially the figurative ones, are radically context dependent, and “the context is not fixed independently of the utterance but constructed as an integral part of the comprehension process” (Wilson and Sperber 2002: 588). We already know this line of argumentation from Carston’s (2002) underdeterminacy thesis. The very notion of ‘what is said’ is misguided from their point of view. Third, Wilson and Sperber argue that Grice has nothing to say about cases like (21), i.e., cases of “loose use”. (21)

a.

The lecture starts at five o’ clock.

[not exactly at five o’ clock]

b.

Holland is flat.

[not exactly flat]

c.

I must run to the bank before it closes.

[not exactly run]

d.

I have a terrible cold. I need a Kleenex.

[not exactly a Kleenex]

Assertoric commitment   

   79

In a strict interpretation of the maxim of truthfulness, these utterances are false. Grice, however, has no strategy to analyze them (they do not fit in his system of violations), but the Relevance-theorists, with their notion of explicature, have. All in all, Wilson and Sperber (2002: 585) concede that “hearers expect to be provided with true information”. But according to them “there is an infinite supply of true information which is not worth attending to. Actual expectations are of relevant information, which (because it is information) is also true.” And for them, it plays no role at all whether this information is literally or conventionally expressed, explicated or implicated. Somewhat surprisingly, Wilson and Sperber (2002: 627) agree with Grice’s statement that “false information is not an inferior kind of information; it just is not information” (Grice 1989d: 371). In contrast, I would like to argue that false information is information indeed: If someone has false information in his common ground because he is lied to, he obviously treats this information as information. This possibility of being deceived about the truth of some proposition shows that the maxim of truthfulness is indeed relevant  – whether in the system of maxims, or in the sincerity or essential condition of assertions (or in both places), remains to be seen. Neither Grice nor the Relevance theorists are concerned with a theory of speech acts, and it is there where lying has to be properly defined in the end.

3.6 Assertoric commitment In (14) above, we have seen that Searle (1979b: 62) proposed a sincerity rule according to which “the speaker commits himself to a belief in the truth of the expressed proposition” and an essential rule according to which “the maker of an assertion commits himself to the truth of the expressed proposition.” It appears, then, that speaker commitment is essential to the definition of an assertion. As MacFarlane (2011: 91) points out, there is a difference to the classic approach of defining assertions in terms of constitutive rules in that the latter approach “looks at ‘upstream’ norms – norms for making assertions”, while “the commitment approach looks at ‘downstream’ norms – the normative effects of making assertions”. I agree that both approaches could be combined (MacFarlane 2011: 91) – in fact, I believe that the proposal of Searle (1979b) in (14) is such a combination. Close to a combined approach is also Alston (2000) because his notion of responsibility comes close to the concept of commitment. MacFarlane (2011: 91) highlights the following four advantages of a commitment approach to assertions over its competitors. First, it gives “a direct description” of what happens if one makes an assertion; second, it can deal with “openly

80   

   Assertion and lying

insincere assertions” which are, according to him, a problem for the expressive approach defended by Bach and Harnish (1979). Third, it can deal with “solitary assertions” and “assertions that cross conversations”, which are problematic for Stalnaker’s approach. Fourth, and most interestingly, the commitment approach is said to offer a natural approach to the speech act of retraction, “as the act of backing out of a commitment to the truth of the asserted proposition”. (This is interesting because there seem to be insincere and untruthful retractions, hence it is possible to lie while retracting an assertion.) Kissine (2008) rightly points out that one finds two views of speaker commitment in the literature: (22)

Two notions of commitment (Kissine 2008) Commitment 1 (Truth-commitment): the speaker, in uttering an assertion, is committed to the truth of p. Commitment 2 (Justification-commitment): the speaker, in uttering an assertion, is committed to provide justifications or evidence for p.

The first view, also dubbed “T-commitment”, is ascribed to Searle (1969) and Alston (2000), the second view, also abbreviated “J-commitment”, is ascribed to Brandom (1994) and Williamson (1996). Note that Grice’s (1989a) maxim of Quality also refers to an obligation to provide evidence for the truth of p. Both types of commitment play a role in asserting. Yet, while J-commitment “boils down to the monotonic commitment to the truth of propositional content” Kissine 2008: 155, and can be cancelled (or retracted, as MacFarlane 2011 puts it), T-commitment is non-monotonic and is associated with all assertive speech acts including “weak” ones. Non-monotonicity goes together with considering alternative possible worlds. “Weak” assertions, such as conjectures or guesses, are often marked with reservation markers or hedges (cf. Overstreet 2011). In English and German, there are several ‘evidentiality-hedges’, as Sweetser (1987: 56) calls them: (23)

a.

to the best of my knowledge; so far as I know; if I’m not mistaken; as far as I can tell; for all I know; as I understand it; my best guess is; speaking conservatively; at a conservative estimate; to put it mildly; beyond question

b.

soweit ich weiß; wenn ich es recht sehe; vereinfacht gesagt; p, sage ich jetzt mal

Assertoric commitment   

   81

Sweetser (1987: 56) points out that “the literal use of these hedges is to limit the speaker’s normal responsibility for the truth of assertions” and adds: “Evidentiality-hedges allow the hearer access to the evidence-evaluation and thus transfer some of the speaker’s evaluative responsibility to the hearer. They avoid potential charges of carelessness or irresponsibility by not allowing the hearer to under- or overevaluate the evidence supporting the hedged assertion.” The question is, then, whether an assertion like As far as I know, there is a traffic congestion on Adenauer Bridge can be understood as the assertion of the speaker that there is a traffic congestion on Adenauer Bridge, and moreover, whether one should assume “half”-, “quasi”-, or “weak” assertions. What is at stake here is the expression of certainty or uncertainty that goes together with an assertion. As I will discuss in section  7.5, most assertions may be qualified according to the degree of certainty that the speaker expresses with respect to their belief in the truth of p. So even if the policeman is uncertain whether there is a traffic congestion on the Adenauer bridge, he nevertheless asserts this. In this respect, consider the possibility of lying while using evidentialityhedges. If the policeman utters As far as I can tell, Smith is the murderer while he has 100 % evidence for this, can he be charged for lying? Or is he only very careful in his wording? These questions lead us deeper into the pragmatic analysis of implicatures because the right answer seems to depend on the right amount of what has been said (as required by the maxim of Quantity). We will come back to this issue in section 4.8. Kissine (2008) basically considers a set of utterances and points out the worries arising from different approaches to asserting, especially the ones by Brandom (1994) and Williamson (1996): (24)

a.

John is a liar. I can’t tell you what makes me think so, but I’d bet my life on it.

b.

The weather is very nice today! [while it is pouring]

c.

The unconscious is structured like a language.

d.

God is everywhere.

For an inferential approach like Brandom’s, assuming (a) “that every successfully asserted content follows from independent justifications”, and (b) “that any such content is a potential justification for further assertions whose contents follow from it” (Kissine 2008: 157), utterances (24a, b) pose problems: The first utterance in (24a) is problematic because the context makes clear that the speaker considers their utterance as unjustifiable, the second, as a case of irony, is false

82   

   Assertion and lying

when read literally. At least these two cases intuitively are assertions yet have to be treated as exceptions in a framework like Brandom’s. The latter problem also occurs with lying; the explanation that the liar only pretends to assert is dismissed by Kissine (2008: 159) because “pretending to assert does not involve any commitment to the truth of the propositional content”. I do not think that this conclusion is right, because in the case of irony the hearer should detect nonliterality, while the pretence-character of a lie is masked. So the duped hearer is licensed to draw inferences from the lie, while the liar might be going on to justify their lie. With respect to Williamson’s knowledge-based approach that contains the constitutive rule ‘Assert that p only if you know that p’, utterances like (24 c, d) are problematic because the contents cannot be known, i.e., there is no demonstrative evidence for them available. Nevertheless, it seems possible to reason about the truth of these sentences, for instance by enlarging the context necessary for their interpretation. So, even in this case, the speaker will be J-committed. Under the heading of “weaker” speech acts, Kissine mentions asserting with reservation, conjecturing, or guessing. These speech acts are not typically associated with J-commitment because the hedges or reservation markers indicate that the speaker has not enough evidence with respect to the propositional content of their utterance: (25)

a.

John was probably there.

b.

I guess/gather that John was there.

However, the speaker is still T-committed. In these cases, the utterances contain two propositions. For instance, in (25a) we have p = ‘John was there’ and ‘the truth of p is probable’. According to Kissine (2008: 164), all assertions are connected with T-commitment, but some assertions are not “demonstratively justifiable”, cf. (24a, c, d). Hence, according to Kissine (2008: 171), “it is not the case that every flatout assertion imposes J-commitment on S. However, in asserting p, S becomes responsible for the persistent truth of p.” With respect to the examples given in (24), I assume that (24a, b) are well-suited for lying, while this is difficult, if not impossible, with (24c, d).

Truth, truth conditions, and truthfulness   

   83

3.7 Truth, truth conditions, and truthfulness In the minimalist model, there is the assumption of a semantic level of truth conditions. Declarative sentences have a connection to truth, and assertions, as types of speech act typically (i.e., directly) conveyed through declarative sentences, do so, too. Often, sentence mood is seen as a semantic property of sentence types that mediates between the declarative sentence associated with certain truth conditions and the assertion in which the speaker is required to have evidence (reasons, etc.) for the truth of p (cf. Searle’s first preparatory rule), or in short, has to be able to justify their assertion. In this section, we will first give a sketch on these different approaches to truth, and then go on to the notion of truthfulness that is important for a definition of lying. A cautionary remark is in order: We will not be able to deal with the notion of truth in any depth because this would lead us into a number of sophisticated philosophical approaches (see the discussions in Künne 2003, Pagin 2007, §5). Instead, we will focus on a notion of truth that is implemented in models of the semantics-pragmatics interface or plays a role in such models. That declarative sentences are connected to truth is captured in the “Most Certain Principle” rendered in Zimmermann (2011: 763) as follows (cf. Cresswell 1982: 69): (26)

Most certain principle If S1 and S2 are declarative sentences such that, under given circumstances, S1 is true whereas S2 is not, then S1 and S2 differ in meaning.

“True” and “false” are truth values; the circumstances under which a declarative sentence is judged to be true (or false) are its truth conditions. The crucial idea in truth-conditional semantics is that a declarative sentence is true when it corresponds to the facts of the world (or, if its truth conditions are fulfilled). Thus, the sentence There is a traffic congestion on Adenauer Bridge is true in case that there is a traffic congestion on Adenauer Bridge. If there is not, the sentence is false. Most importantly, the competence to give such judgments is strongly connected with the knowledge of the words used and the way in which the words are syntactically put together (Frege’s principle of compositionality); if a speaker did not have this competence, they would rarely arrive at the correct results. A further semantic principle mentioned by Zimmermann (2011: 763), who traces it back to Wittgenstein (1922: 4.431), is given in (27):

84   

(27)

   Assertion and lying

Basic principle of truth-conditional semantics Any two declarative sentences agree in meaning just in case they agree in their truth conditions.

From the point of view of semantics, this amounts to the following basic assumption: “Whatever the meanings of (declarative) sentences may be, then, they must somehow determine their truth conditions, i.e. the circumstances under which the sentences are true.” (Zimmermann 2011: 763) What is not captured by truth conditions is left to pragmatics. Problems arise when the notion of a proposition is studied more carefully. Arguably, sentences understood as purely linguistic objects may have truth conditions (in relation to some model world), but it is utterances that are the bearers of truth values. Therefore, Davidson (1984: 135) proposed for (28a) the T-sentence condition (28b) (with “T” for truth or the philosopher Alfred Tarski who invented that schema): (28)

a.

It is raining.

b.

‘It is raining’ is true as uttered by a speaker S at a time t if, and only if, it is raining near S at t.

(28b) captures the fact that every utterance is related to a particular speaker, some point in time when the utterance is made, and a certain place where the speaker is situated at the moment of speaking. The need for taking these variables into account for a speaker, a time, and a place is most clearly seen with deictic expressions. Take (29), which without any doubt is a sentence, as an example: (29)

I am here now.

Thus, deictic information must be known before the truth of the sentence may be evaluated. The point is that this information is only given in a concrete situation of utterance. For instance, (29) could be understood as Boris Becker is at Mainz University at December 5th, 2008. Independent of such a concrete situation, I means “the speaker”, here the place of the speaker, and now the time of speaking. Indexical Pragmatics, as Levinson (2000) dubs it, is a component that determines truth conditions, but must have access to the context of an utterance. This, however, was already accepted in the classic Gricean model with its opposition between ‘what is said’ and ‘what is implicated’.

Truth, truth conditions, and truthfulness   

   85

As for the completeness of propositions, recall the short discussion on the underdeterminacy of propositions. Standard examples (beyond the resolution of ambiguities, indexicality assignment, and cases of ellipsis like On the top shelve as an answer to a preceding question) include the case of missing constituents, unspecified scope, underspecificity or weakness of constituents, and overspecifity or narrowness of constituents (Carston 2002; see (2) in section 2.2). For example, in an utterance like Paracetamol is better [than what?], there is a constituent missing that must be supplied pragmatically (Carston 2002: 22). Only by completion of the utterance it is possible to judge it as true or false. Note that Carston (2002: 22) also adds the example It is raining that we already know from Davidson’s account. This nicely points out the basic question how we arrive at completing the underdetermined proposition. We will come back to this question in section 4.7. Just as in truth-conditional semantics the meaning of a declarative sentence is identified with its truth conditions and truth values, the pragmatic definition of assertion is also related to truth. Pagin (2007) mentions several conceptual strategies how to bring assertion and truth together, some of which we already have come across. The first strategy, called “assertion as truth-claim”, is to assume that in asserting, the speaker presents a proposition as true. In this sense, the speaker of an assertion is “claiming” that a proposition is true. The second and third strategies both consider truth as an aim of the speaker’s assertion. While the second strategy, assertion as fact-stating, involves the idea that the speaker aims at representing the facts, the third strategy invokes a norm or convention that in asserting, the speaker aims at saying something that is true. The Gricean maxim of Quality may be seen as such a norm (Grice 1989a). The fourth strategy has to do with correctness, where correctness is understood as some kind of measure for the degree of truth. Thus, if a speaker has good evidence for p, then their assertion may be evaluated as correct. The fifth strategy is to bring together truth and belief, or truth and knowledge, respectively. We know the former approach from Searle (1969) and Bach and Harnish (1979), the latter approach from Williamson (1996). What this survey of approaches to the role of truth in defining assertion suggests, is that the notion of truth, when applied to assertion, is by no means selfexplaining. It appears that all of the mentioned strategies have a certain point. Looking back at the truth-conditional approach, we may suspect that while the connection of the meaning of a declarative sentence to the truth of a proposition is intuitively clear, the exact meaning of p “being true” (with respect to specific circumstances) is not. As already mentioned, in several approaches there is an intermediating level between the declarative sentence type and the illocutionary force of assertion,

86   

   Assertion and lying

namely sentence mood or sentence modality. In these modular approaches that combine generative syntax with truth-conditional semantics, the semantics of declarative sentences, echoing Searle’s classic formula “F(p)”, contain an operator ASS that takes the proposition of the declarative sentence in its scope. Other approaches assume that the existential operator is somehow connected to the sentence mood of declarative sentences (e.g., Brandt et al. 1992). And there are even influential approaches that incorporate a FORCE-projection into sentence syntax, e.g. Rizzi (1997). The basic idea behind these frameworks is certainly that sentence structures encode assertivity. Yet it is not always clear what the suggested relation between assertion and truth is. (In the next section, we will go into one influential semantic approach to declarative mood, the one proposed by Robert Stalnaker and his followers.) We continue with a short discussion of the notion of truthfulness which is relevant for the topic of lying. As we have already seen, when one speaks about the truth or falsity of some state of affairs, it may be asked how we can be sure what the truth is. I refrain here from going into this epistemological question in any detail but would like to assume that there exists something like an objective state of affairs, as Dietz (2002: 75) puts it, and that this state may be judged as true or false: “Objektive Sachverhalte sind empirisch wahrnehmbare, raumzeitlich datierbare Ereignisse oder komplexe Zustände, die als solche grundsätzlich für jedes Erkenntnissubjekt in gleicher Weise zugänglich und überprüfbar sein können.” ‘Objective states of affairs are empirically perceivable events (located in space and time) or complex states which basically are accessible and provable for every subject striving for knowledge in the same manner.’ [My translation, JM]

Restrictions are due to the historicity of states of affairs (e.g., historical states of affairs cannot be testified), spatial distance to the state of affairs, lack of cognitive competence, or lack of means that could verify or falsify a certain state of affairs. Dietz (2002: 75) points out that a precondition for a truth judgment is that there is a clear reference to some state of affairs, that it can be negated, that it is without contradiction (Germ. Widerspruchsfreiheit), and that it can be proved. In short, empirical utterances must be falsifiable. This is quite straightforward. Why, then, do we need a further notion, the notion of truthfulness (Germ. Wahrhaftigkeit)? While the notion of truth aims at an objective, verifiable aspect of judging a certain state of affairs, the notion of truthfulness aims at the subjective aspect, i.e., what the speaker believes to be true (Falkenberg 1982: 54–59; Dietz 2002: 64–81; see also Williams 2002, Vincent 2006). A popular or classical view of lying defines every false utterance as a lie.

Truth, truth conditions, and truthfulness   

   87

But first, it is possible to lie while using true utterances, and second, it is possible to be honest (truthful) while using false utterances (mistake). Let us illustrate the different instances of being or not being truthful with the example of a speaker S, called Barbie, who utters I’m suffering from a heart disease (cf. Falkenberg, 1982: 54–58, building on an example by E. Utitz). First, Barbie believes at t that she is suffering from a heart disease and it is indeed the case that she is suffering from a heart disease. The propositional content p is true and the speaker is truthful (she did not lie). Second, Barbie believes at t that she is suffering from a heart disease but, actually, she is not. Thus, she is mistaken. The propositional content p is false, but the speaker is truthful (she did not lie). Third, Barbie does not believe at t that she is suffering from a heart disease, and she knows that, as a matter of fact, she is not suffering from a heart disease. It follows that she is lying (she is not truthful). Fourth, Barbie does not believe at t that she is suffering from a heart disease, but in fact she is. She is mistaken since the propositional content p is true. Moreover, she lies (she is not truthful) since she wants to deceive. For a theory of lying it is very important that it can deal with this last case. With regard to the speaker’s intention to deceive, it does not matter what actually is the case; it only matters what the speaker believes to be the case. Note that this assumption is still open to dispute. As far as I can see, the prevalent opinion is reflected in the following quote from Davidson (1984: 258): “It is sometimes said that telling a lie entails what is false; but this is wrong. Telling a lie requires not what you say be false but what you think is false. Since we usually believe true sentences and disbelieve false, most lies are falsehoods; but in any particular case, this is an accident.”

Rott (2003: 10) holds, in contrast, that a true assertion cannot count as lying. This is shown by the following dialogue: (30)

[Person A utters a true sentence, but believes that it is false] B: Warum hast du mich angelogen? ‛Why did you lie to me?’ A: Aber was ich sagte, war doch wahr! ‛But it was true what I said!’ B: Ja, aber du wolltest mich doch anlügen! ‛Yeah, but you wanted to lie to me!’

I think, however, that this example only shows that, as a reaction to a reproach, it is possible to advocate the popular position that lying essentially consists of

88   

   Assertion and lying

saying something false. When the attempt at deception is included, no matter whether the propositional content of the lie is accidentally true, the reproach of B certainly is to the point. In sum, then, I assume that we need two levels of analysis: truth conditions and truth values on the one hand, being associated with the level of truth-conditional semantics, and the level of truthfulness on the other hand, being related to inner mental states of speakers. This aspect is not easily captured by standard minimalist models, but it is presupposed insofar as speakers are usually assumed to have certain belief states towards a proposition.

3.8 Assertion, common ground, and presupposition In many treatments of assertions it is assumed that they change the common ground. The notion of common ground is strongly connected with the work of Stalnaker (1979, 2002) and his theory of presupposition. The basic idea is this: In a dialogue, it is very important to keep record of the contents that are either believed by only the speaker S or only the hearer H, and most importantly, those contents that are believed by both of them. The common ground is then the set of propositions that are believed by both the speaker and the hearer. These propositions constitute the respective ‘belief sets’ (BS) of the speaker and the hearer. In a set-theoretic representation, we may write: (31)

BSS BSH

= =

{p: S believes p} {p: H believes p}

Consequently, the common ground may be represented as the intersection of the two belief sets. There could be a situation in which both speaker and hearer have the same proposition in their belief sets without knowing that the other shares this believed proposition. Therefore, it is crucial to define the common ground in a way that the definition reflects that the respective speakers know that the other has a certain believed proposition in his belief set. In short, the propositions in the common ground must be mutually known by both speaker and hearer. Hence we define: (32)

CG{S, H} = {p: p ∈ BSS & p ∈ BSH & p is mutually believed by both S and H}

A partial set of these propositions will be true with respect to a possible world. This set of true propositions is called the ‘context set’.

Assertion, common ground, and presupposition   

   89

It follows that propositions may have different statuses. Thus, it could be the case that there are private propositions for speaker and hearer, i.e., propositions that are not shared; or that there are accidentally shared propositions, i.e., speaker and hearer do not know that they share a proposition; or that there are shared propositions mutually known to speaker and hearer. With every new assertion, the common ground may be enlarged. Of course it is possible to reject an assertion. If it is rejected, it will not be part of the common ground. False propositions may be part of the common ground if they are mutually believed by the speaker and the hearer. But lies can never be part of the common ground since for lies it is constitutive that the speaker does not believe a certain proposition, while he behaves as if this were the case. Thus, the deceived hearer may have the false proposition contained in a lie in his belief set, but the liar does not. Note that Gunlogson (2001: 42) proposes the notion of ‘public belief’ according to which it is mutually known that someone has a certain belief. (33)

S and H mutually believe q = ‘H believes p’

If p is false and is taken over from S, then S can certainly control the success of their lie. In this case, H would surely believe that S believes p, too, but in fact, S does not. The common-ground approach allows for the inclusion of presuppositions in that it assumes that presuppositions are propositions that already are in the common ground. Logical entailments, so it could be argued, as well as implicated propositions (if shared), also form part of the common ground. As Pagin (2007: 3) points out, “assertion is generally thought of as being open, explicit, and direct, as opposed to conveying something indirectly, without explicitly saying it.” On the other hand, assertions are connected to presuppositions and implicatures, and they may be conveyed indirectly, as indirect speech acts. In this section, we will sketch some aspects of the relation between assertion and presupposition. Basically, presuppositions are additional propositions. In this respect, they resemble implicatures. There is a long-standing debate about the question whether presuppositions are semantic or pragmatic in nature. The semantic line of argument points to the semantics of so-called presupposition triggers, such as factive verbs, definite descriptions, lexical elements like to manage, to begin, again (Levinson 1983: §  4.2), wh-interrogatives, non-restrictive relative clauses. Traditionally, the presupposition behavior of such items is tested with the negation test, requiring constancy under negation. The pragmatic line of argument points to the defeasibility of presuppositions in some contexts. If presuppositions

90   

   Assertion and lying

were simply a special type of entailment (p→q and p→ not q), defeasibility should be impossible. Chierchia and McConnell-Ginet (1990: 23–28, 280–295) propose the S-familytest (with “S” for sentence). They observe that a presupposition does not only show up in a positive sentence and its negative counterpart, but also in other sentence types: (34)

a.

Ken has quit dealing.

b.

Ken has not quit dealing.

c.

Has Ken quit dealing?

d.

If Ken had quit dealing, Barbie would be glad.

All of these sentences presuppose (and this has to do with the verb to quit s.th.) that Ken has been dealing. Chierchia and McConnell-Ginet assume that this information is part of the background knowledge of interlocutors. Presupposition-triggering elements thus serve to make these background assumptions clear in the course of the ongoing discourse. As mentioned in the previous section, Stalnaker’s approach to the common ground provides a particular division of conversational labor: While assertions go into the belief set and are mutually known by the discourse participants, presuppositions, as made visible by the so-called presupposition triggers, are propositions that are already in the belief sets. The main problem with this theory is that presuppositional constructions might introduce new information (cf. Jary 2010: 92–99). For instance, with the utterance of Ken gave his aunt roses we may introduce the information that Ken has an aunt. Stalnaker attempts to solve this problem by making use of the notion of accommodation. Jary (2010: 95) explains his idea as follows: “When a linguistic form that is generally only appropriately used if P is common ground, then the participants of the conversation tend to accommodate that use by quietly adding p to the common ground.” (See also Beaver and Zeevat 2007.) More elaborated, this amounts to a reasoning process of the hearer along the following lines (Jary 2010: 98): (35)

1.

S’s utterance contains L.

2.

It is communicatively efficient to utter L only if P is common ground.

3.

P is not common ground.

4.

S is not behaving in a communicatively efficient manner,

Indirect assertions and explicit performatives   

   91

5.

If the hearer accommodates P, S will be behaving in a communicatively efficient manner.

6.

P is not controversial and S is in a position to know that P.

7.

Accommodate P!

It is clear, as Jary emphasizes, that the premise in line 5 contradicts the premise in line 2. Hence, he concludes that this move in Stalnaker’s theory is not plausible since the phenomenon of accommodation is so widespread that it appears to render the concept of presupposition vacuous. Similar criticisms have been put forward by Abbott (2000), Simons (2003, 2005), and Zhang (2008). Thus, while the notion of accommodation appears to be problematic, the dynamical approach to the exchange of assertions and the roles presuppositions play in this respect is relevant for the enterprise of fixing lying at the semanticspragmatics interface. Minimalist models like the ones by Recanati and Levinson have no systematic dynamic aspect. However, when it comes to modelling real talk exchange with respect to common ground, minimalist models have to be enriched accordingly.

3.9 Indirect assertions and explicit performatives Definitions of assertions are usually meant to cover standard cases of speechact exchange in which the speakers are sincere, the words used have their literal meaning, the situation of the utterance is neutral, etc. In short, the illocutionary force indicating devices (IFID) contained in a sentence (or sentence fragment) are sufficient to signify a certain type of speech act (illocution). Thus I promise to come is a promise in virtue of the performative verb to promise, and Do you like to come? counts as a yes/no question in virtue of the interrogative sentence. But there are problems with this approach, as can be seen in the examples in (36): (36)

a.

Boy, I promise to pay that back to you!

b.

Who in the world will ever be interested in this bullshit?

Despite of the verb to promise in (36a) and the wh-interrogative in (36b), these utterances are best interpreted as a threat and as an assertion. This kind of nondirect or “indirect” interpretation of a speech act poses a serious problem for most accounts of the semantics-pragmatics divide because it has to be explained

92   

   Assertion and lying

why certain semantics connected with a sentence type (or sentence mood) and a specific lexical make-up get inactive. Searle (1979a) proposes a strategy how to solve the problem. The basic idea is very simple. Because the illocutionary force indicator obviously does not signify the illocution it should indicate, some other illocution must be meant. Taking the whole situation of utterance into account, the hearer arrives at a plausible hypothesis about what that illocution might be. This inferencing process is similar to the inferencing schema for implicatures proposed in Grice (1989a). Most importantly, the inference involved is a “practical” inference, one, that proceeds by hypothesis formation, not by logical entailment. In Meibauer (1986), I propose to analyze rhetorical questions as indirect assertions. Moreover, I identify rhetorical commands and rhetorical assertions. The following table provides some examples: (37)

Rhetorical questions, commands, and assertions

IFID – sentence type Am I responsible for Yes/no interrogative Wh-interrogative thist disaster?Who wants that? Show me one person imperative that likes Bush! I want to see the person that likes Bush!

declarative

Intended reading ‘I am not responsible for this disaster.’ ‘Nobody wants that.’ ‘You are not able to show me one person that likes Bush!’ ‘There is no person that likes Bush.’

Illocution Assertion

Assertion

Assertion

If this is a plausible analysis, we are faced with an interesting question: Is it possible to lie while using a rhetorical question? I would like to answer the question with “yes”. It is easy to see why. If someone is in fact responsible for a particular disaster (because, in fact, they caused it), and, in the course of a meeting, rhetorically utters Am I responsible for that disaster?, thereby conveying the assertion ‘I am not responsible for this disaster’, then we can say about this speaker that he or she is a liar. As is well-known, it is not possible to use the verb to lie in an explicit performative construction. To say I hereby lie that p with the intention to make the insincere assertion explicit, would lead to an “illocutionary suicide”. We know that the successful liar masks their assertion as being true, as being conform to the conventions associated with honest assertions.

Indirect assertions and explicit performatives   

   93

Since we want to associate the declarative sentence with assertions, we have to account for an old problem in the analysis of explicit performatives. (38)

a.

I (hereby) promise that I will come to the party.

b.

I (?hereby) assert that Ken came to the party.

c.

Ken came to the party.

Because of the prototypical illocutionary force indicating device to promise, (38a) is taken as a direct promise in classical speech-act theory (Searle 1969). Since (38a) displays a declarative sentence that is also an illocutionary force indicating device signalling assertions, we have a conflict here. Therefore, Bach (1975) and Bach and Harnish (1992) argue that “performatives are statements too”, i.e., all explicit performative constructions are indirect speech acts in which the secondary illocution is an assertion (statement) and the primary speech act is the one indicated by the performative verb. They criticize Searle’s (1989) idea that explicit performatives are still direct speech acts because of their self-guaranteeing character. According to Searle, every explicit performative utterance is a declaration. Thus, the utterance (38a) has the effect of bringing about a state in which the promise is a fact. Moreover, the performative is also a statement and has, of course, the force denoted by the performative verb. Thus, his analysis yields that the utterance I hereby order you to go comprises the elements of a directive, assertive, and declarative act without any recourse to indirectness. Bach and Harnish (1992) deny that all performative constructions have an inherent self-guaranteeing character and that they are declarations. The selfguaranteeing is as best a default since it can be blocked by other information, for instance, specific gestures. Only a subset of declarations, e.g. blessings, are declarations. And finally, Searle has no explanation at hand for why statements should play a role in the analysis at all (see also Reimer 1995). Now it seems possible to use the verb to assert in an explicit performative construction, as in (38b). Here, the force indicated by the performative verb and the declarative sentence collapse. However, (38b) is not equivalent in function to (38c). While (38c) has the force of an assertion, (38b) has the force of insisting on an assertion. Note in addition that the insertion of the indicator of tokenreflexivity hereby is somewhat strange. So the utterance in (38b) requires very special contexts. Is it possible to lie by using indirect speech acts and explicit performatives? My intuition is that this is possible. Imagine Ken, the radical pragmaticist. Invited by the influential Society for Hardcore Semantics, he ends his speech with the rhetorical question Who on earth needs contextualism? His audience takes this

94   

   Assertion and lying

utterance as conveying the indirect assertion ‘Nobody needs contextualism’. But this is, of course, not Ken’s true belief, so what he did was lying to his audience. As for explicit performatives, consider the utterance I hereby assert that Bush was honest, spoken by Ken who works for Bush in the context of a TV discussion over the question of Bush’s honesty. Ken’s utterance is a clear-cut lie if he knows that Bush was dishonest. This shows that it is possible to lie by using explicit performatives.

3.10 Expressive meaning Besides the propositional meaning of a sentence which is related to truth conditions (or answer/fulfilment conditions, Boisvert and Ludwig 2006) and the meaning phenomena of entailment, presupposition, and implicature, there is at least one further aspect of meaning which is relevant on the level of sentence meaning, and this is expressive meaning. In this section, I will first point out Kaplan’s view of expressive meaning before I will discuss Potts’ framework. Finally, I will consider the relation between conventional implicature and expressive meaning. At first glance, we may characterize expressive meaning as something which represents a certain emotional state of the speaker. This is indeed the view of Potts (2007: 173) when he states: “Almost invariably, a speaker’s expressives indicate that she is in a heightened emotional state.” While this may be correct in many cases, I do not think that this is a necessary condition for expressiveness since clear-cut ethnic slurs may be used expressively without any “heightened emotional state” of the speaker. Kaplan’s (1999: 5) program amounts to capturing expressive meaning within a system of formal semantics: “So here is my method. I ask not what the expression means (for example, I do not ask, ‘What does “Goodbye” mean?’). Instead, I ask, ‘What are the conditions under which the expression is correctly or accurately used?’ This seems a much more fruitful line of inquiry for a word like ‘Goodbye’. To the degree that such conditions reflect linguistic convention, the information that such conditions obtain is carried in the semantics of the expression.”

A standard view of the meaning of expressive items like damn is, according to Kaplan (1999: 7) that “logic (and perhaps even truth) is immune to epithetical color”. In contrast, Kaplan (1999: 6) argues that “for certain expressions of natural language, a correct Semantic Theory would state rules of use rather than something like a concept expressed.” Hence he assumes that the sentence The damn Kaplan was promoted is true if Kaplan was promoted and the agent has

Expressive meaning   

   95

a derogatory attitude towards him (Kaplan 1999: 10). (Note that McCready 2012 argues that damn and fucking may convey negative as well as positive emotional states.) The same analysis applies to interjections such as ouch. If an agent in a certain context has just experienced a sudden pain and utters Ouch!, the semantic information will be the same as if he said I am in pain. These expressions are “informationally equivalent” (Kaplan 1999: 16); however, Kaplan hastens to add that this does not include “what can be derived from the fact that someone has used one of them” (Kaplan 1999: 17). One wonders how one has access to the informational level when abstracting away from the context of use. If the test of commenting with That’s true! is applied, it is clear that one cannot react this way to Ouch!, while reacting like this to I am in pain is possible. To integrate use conditions into semantics appears to be a move towards blurring the traditional distinction between semantics and pragmatics. More recently Potts (2007) has developed a heuristics for the expressive dimension. Expressiveness has to do with what he calls conventional implicature. We will first go into his criteria before we consider the notion of ‘conventional implicature’. The first criterion is Independence, i.e., an expressive content can be separated from regular descriptive content. In particular, the independence property says “that we can change or remove the expressive content of a phrase without affecting its descriptive content” (Potts 2007: 168). This is demonstrated by the expressive apposition bastard: (39)

a.

That bastard Kresge is famous.

b.

Descriptive: Kresge is famous.

c.

Expressive: Kresge is a {bastard/bad in the speaker’s opinion}.

The second criterion is called Nondisplaceability and says that expressives predicate something about the utterance situation, and hence “cannot (outside of direct quotation) be used to report on past events, attitudes, or emotions, nor can they express mere possibilities, conjectures, or suppositions. They always tell us something about the utterance situation itself.” (Potts 2007: 169) (40)

a.

That bastard Kresge isn’t late for work. (#He’s a good guy.)

b.

It’s just false that that bastard Kresge is late for work (#He’s a good guy.)

96   

   Assertion and lying

c.

#If that bastard Kresge arrives on time, he should be fired for being so mean.

d.

Maybe that bastard Kresge will be late again. (#Then again, maybe he’s not a bastard.)

However, there are problematic cases in which a certain perspective is built into a speech report, e.g., My father screamed that he would never allow me to marry that bastard Webster (Kratzer 1999) where it is clear that the father thinks that Webster is a bastard. This observation leads to the third criterion of Perspective dependence, meaning that expressive content is evaluated from a particular perspective. In general, the perspective is the speaker’s, but there can be deviations if conditions are right. Potts (2007) thinks of introducing a so-called “judge” here, understood as someone who holds a certain evaluation, for example, the father in the above example. However, for Potts (2007: 176) “it appears to be a marked option to evaluate expressives with a judge who is not also the speaker […]”. The fourth criterion is Descriptive ineffability, i.e., speakers are never fully satisfied when they paraphrase expressive content by using descriptive content. For instance, it is not obvious that ‘Kresge is bad’ is an adequate paraphrase for that bastard Kresge, see (39c). The fifth criterion, immediacy, tells us that “expressives achieve their intended act simply by being uttered” (p. 167); in this respect, they are similar to performatives. The sixth criterion is Repeatibility, i.e., the speaker may use an expressive item repeatedly and thereby strengthen the emotive content. This is not possible with descriptive content where repeated content leads to redundancy. While this list of criteria is suggestive, it goes without saying that the criteria are somewhat intuitive. For instance, “immediacy” seems to be connected to all speech acts, whether they are connected to expressiveness or not. Moreover, there might be further criteria, e.g., expressives cannot be coordinated (??that bastard and idiot Kresge) and cannot be negated (*that non-bastard Kresge). Since our discussion is motivated by the question whether it is possible to lie while using expressives, for instance, if the speaker admires Kresge at the time they utter (39a), we should know what the exact meaning status of expressive elements is. Basically, there are several options. First, we can deal with expressive content as if it were regular descriptive content; then bastard has as part of its literal meaning a pejorative content. (This approach was already dismissed by Frege in his discussion of the meaning of Köter ‘cur’ in distinction from Hund ‘dog’.) Second, we can deal with expressive content as a conventional implicature, as Potts does. (Note that Horn 2008 speaks of an F-implicature here,

Expressive meaning   

   97

with “F” for “Frege”.) Third, we can argue that expressive meaning is presuppositional in nature, as Schlenker (2007) does. Let us concentrate on the second approach. Recall that Grice (1989a) is partly motivated by rescuing the logical meaning of and as part of natural language meaning. In sentences like He is an Englishman; therefore, he is brave, the expression therefore basically behaves like the conjunction and but contributes a conventional implicature like ‘His braveness is a consequence of his being an Englishman.’ Conventional implicatures are by definition not cancellable, reconstruable, and context-variable. They are conventional, because they are not conversational; they are implicatures, because they do not belong to the realm of ‘what is said’. Now Bach (1999) argues that ‘conventional implicature’ is a superfluous concept because it assumes the unwarranted principle “one sentence  – one proposition’ (OSOP, for short). Instead, Bach assumes that a complex sentence like Gabriella is rich but happy has three propositions: (a) that Gabriella is rich, (b) that she is happy, (c) that there is a contrast between being rich and being happy. For Bach the proof for this multipropositional approach is that in speech reports like Marv said that Shaq is huge but that he is agile the contrastive meaning is part of what Marv said, and hence part of ‘what is said’ (cf. Conrad 2011). In Potts approach, ‘conventional implicature’ gets a new outfit. He defines the term as follows: (41)

Conventional Implicature (Potts 2005: 11) a. CIs are part of the conventional meaning of words. b.

CIs are commitments, and thus give rise to entailments.

c.

These commitments are made by the speaker of the utterance ‘by virtue of the meaning of’ the words he chooses.

d.

CIs are logically and compositionally independent of what is ‘said (in the favored sense)’, i.e. independent of the at-issue entailments.

According to Potts (2005), supplements like nominal appositions, non-restrictive relative clauses, and expressives like damn, that bastard are conventional implicatures, i.e., the definition in (41) applies (cf. Amaral, Roberts, and Smith 2007, Jary 2010: 134–137, McCready 2010). As far as I see, most researchers would agree that there is non-propositional meaning (see the meanings of interjections or German modal particles) but it is still controversial whether the concept of conventional implicature as in (41) is needed.

98   

   Assertion and lying

A quick look at German data shows that there are many elements that do no contribute to truth conditions but modify or qualify the content of the utterance. Here are some typical examples connected to the expression of regret (MP = modal particle): (42)

a.

Es gibt halt einen Stau auf der Adenauer-Brücke. ‘There is MP a traffic congestion on the Adenauer bridge.’

b.

Leider gibt es einen Stau auf der Adenauer-Brücke. ‘Unfortunately, there is a traffic congestion on Adenauer bridge.’

c.

Bedauerlicherweise gibt es einen Stau auf der Adenauer-Brücke. ‘Unfortunately, there is a traffic congestion on Adenauer bridge.’

d.

Es ist zu bedauern, dass es einen Stau auf der Adenauer-Brücke gibt. ‘It is unfortunate that there is a traffic congestion on Adenauer bridge.

e.

Ich bedauere, dass es einen Stau auf der Adenauer-Brücke gibt. ‘I am sorry that there is a traffic congestion on Adenauer bridge.’

f.

Mit dem Ausdruck des Bedauerns stelle ich fest, dass es einen Stau auf der Adenauer-Brücke gibt. ‘With an expression of regret I state that there is a traffic congestion on Adenauer bridge.’

All the elements in italics serve the purpose to express that the speaker is sorry that p. The question is whether these elements contribute to the truth or falsity of the propositional content or not. Most researchers make a distinction between the non-propositional and the propositional expression of psychic attitudes. Thus, the expression of regret is non-propositional in (42a–c), but propositional in (42d–e). Our question is whether it is possible to lie with the utterances under (42). First, it could be argued that that is impossible because inner states cannot be evaluated in terms of truth and falsity in any way. Second, it could be argued that is possible to lie only with (42d–f) because the attitude is propositionally expressed in these cases, in sharp contrast to (42a–c). Third, it could be argued that in all these cases lying is possible if the speaker does not have the attitude they pretend to have. Whether this can be controlled by others is of secondary interest. Dietz (2002: 74), as far as I see, opts for the latter position. She argues that the speaker is in the privileged position of knowing whether their attitudes are true or not:

A simple definition of assertion   

   99

“Der besondere Fall der Lüge zeigt, daß es durchaus sinnvoll sein kann, bezüglich psychologischer Sachverhalte von einem Wissen zu sprechen: Wer lügt, daß er eine bestimmte Wahrnehmung oder ein bestimmtes Gefühl nicht hätte, weiß gleichzeitig, dass er diese Wahrnehmung oder dieses Gefühl hat.” [‘The special case of a lie shows that it can be absolutely reasonable to speak of knowledge with respect to psychological states: They who lie that they do not have a specific perception or a specific emotion at the same time also know that they do have this perception or this emotion.’] [My translation, JM]

If this is on the right track, we have to conclude that propositional expression is not a precondition for lying and that lying is also possible with the expression of false attitudes, whether they are propositionally or non-propositionally expressed. Alternatively, we would speak of deceiving instead of lying.

3.11 A simple definition of assertion As we have seen in the foregoing sections, there are many perspectives on how to characterize an assertion. Recent research such as Pagin (2007), Jary (2010), Brown and Cappelen (2011) shows that assertion is not a simple concept, as is often assumed. So Cappelen (2011) certainly has a point when he argues that eventually a category of assertion will be superfluous because one can reach all descriptive and explanatory aims by assuming the notion of saying plus a correlative notion such as commitment or cause. Yet for our purposes, namely analyzing lying at the semantics-pragmatics interface, this approach is far too deconstructive. Therefore, I will propose a definition of assertion that captures many insights of the preceding discussion, but is not overcomplicated and burdened with special theoretical assumptions. Let us assume that an assertion is always made by a speaker S at time t, and that it has a certain propositional content p. Improving a proposal of Falkenberg (1982: 91), we may then define an assertion as in (43): (43)

Assertion S asserted at t that p iff (a) S uttered at t the declarative sentence σ meaning p, (b) by uttering the declarative sentence σ, S presented p as true, (c) by uttering the declarative sentence σ, S M-intended that an addressee H to whom S uttered p actively believes that p.

I will comment shortly on (43a–c) before I deal with possible problems of this definition.

100   

   Assertion and lying

Ad (43a). Condition (a) requires that a declarative sentence is used. We have already shown that assertions are systematically connected to declarative sentences, and from a linguistic point of view this should be built into a definition of assertion. Since we know that assertions may be conveyed indirectly, e.g., by means of an interrogative sentence, it follows that the definition of assertion given in (43) cannot capture this fact. Definition (43) is not intended to account for notorious problems in the grammar-pragmatics interface such as indirect, non-literal, or underdetermined assertions. To account for these phenomena, mechanisms of modulation or pragmatic inferences are required that lead to interpretations beyond the standard speech-act assignments. Ad (43b). Condition (b) is a condition that draws on the typical assumption to be found in many definitions of assertion, namely that in asserting p, the maker of an assertion commits themselves to the truth of the proposition expressed (e.g., Searle’s essential condition). By presenting p as true, the speaker may also be seen as representing themselves as believing p; as Pagin (2007: 14) summarizes: “By means of answering the question what I believe with an utterance of (33) [=There are black swans, JM] I do represent myself as believing that there are black swans, equivalently with asserting it.” The idea of “representing oneself” by assertion is traced back to Black (1952) and Davidson (1984: 268). Ad (43c). This condition is useful in that it situates assertion within the Gricean theory of communication. A M(eaning)-intention is an intention that is necessary for the speaker’s attempt to produce a certain belief in the addressee by their very utterance. Following Falkenberg (1982: 88), who draws on Grice (1989c) and Schiffer (1972), the M-intention may be defined as follows: “A uttered s with the M-intention to produce the belief that p iff A meantNN p by uttering s”, where meaningnn is the Gricean abbreviation for non-natural (i.e. verbal) meaning. The crucial point is that by asserting p the speaker wants the addressee to share p with them (or to take p in their belief box, as Kissine 2008 puts it). A possible objection to condition (c) draws on the possibility of asserting something “without intending the audience to actively believe it”. A case in point would be that the speaker is knowing in advance that the audience will disagree. According to my analysis, however, every assertion is connected with condition (c) for the simple reason that, in contrast to statements, assertions essentially are a means to persuade the audience that p. Thus, even if the audience is supposed to not believe p, it is nevertheless useful to assert p from the point of view of the speaker.

A simple definition of assertion   

   101

Note in addition that one should make a distinction between conditions of success, conditions of satisfaction, and conditions of acceptance (see Rolf 1997 for an overview). If an assertion is not accepted by the audience, this will be captured by conditions of acceptance. In definition (43), we concentrate on conditions of success only. Falkenberg (1982: 56) assumes that if someone is not lying, they are truthful. By and large this seems right, with two putative exceptions (brought to my attention by Jonathan Adler and Christian Plunze). First, S neither believes that p nor that not p, but asserts that p. This may happen within a particular discourse “for the sake of argument”. However, it seems to me that this is not really an assertion. The propositional content is not so much presented as true by the speaker, but presented as something that could either be believed or disbelieved. Second, S fails to believe p, but asserts p. Again, this behavior, however it may be caused, is defective as an assertion because the speaker only pretends to present p as true. In summary, then, the following picture of the formation of assertion arises. The linguistic basis of a prototypical assertion is the declarative sentence that is grammatically endcoded: verb-second with [-wh] in the prefield plus falling intonation. The lexical make-up of the declarative sentence determines its propositional content that has to be enriched in most cases, be it on the indexical level or with respect to underdeterminacy. If we have a complete proposition, then the truth conditions of the declarative sentence may be calculated. After this is done, we have to account for presuppositions, conversational implicatures and speech act assignment. Finally, an assertion is embedded into discourse, i.e., it is connected with mechanisms of determining a common ground and fitting moves in the sequencing of utterances. In the next chapter, I will propose a definition of lying that fits the definition of assertion proposed here, and that acknowledges the possibility of lying while saying the truth.

4 Defining lying 4.1 Introduction There are numerous proposals in the literature of how to define lying. Instead of comparing the different proposals and pointing out their weak and strong aspects, I will present a reasonable definition of lying, and then point out some critical aspects. The first problem is whether lying presupposes an intention to deceive. Pace Fallis (2009) and Carson (2010), I will argue that this is indeed the case, although it is not necessary to include this aspect directly into the definition of lying. The second problem is whether the proposed definition is able to account for self-deception. I will argue that this is a special case that does not need to be covered by the definition of lying. In a next step, drawing on earlier work (Meibauer 2005, 2011), I will propose to extend the definition of lying so that untruthful implicatures are part of the act of lying. Elaborating on this approach, I will go on and ask whether untruthful presuppositions and – if this category is acknowledged – untruthful conventional implicatures should be included in the extended definition of lying, too. Since explicature or impliciture are extensions of the propositional content, it should also be possible to lie while using untruthful explicatures or implicitures. Finally, I will show that phenomena of vagueness and imprecision invite the liar to make use of them, thus reducing their risk to be caught in the act.

4.2 A definition of lying Recall the definition of assertion we gave in section 3.11, here repeated for convenience: (1)

Assertion S asserted at t that p, if and only if (a) S uttered at t the declarative sentence σ meaning p, (b) by uttering the declarative sentence σ, S presented p as true, (c) by uttering the declarative sentence σ, S M-intended that an addressee H to whom S uttered p actively believes that p.

On the basis of definition (1), we may define the term ‘lying’, roughly following Falkenberg’s approach, as in (2) (cf. Falkenberg 1982: 75):

A definition of lying   

(2)

   103

Lying S lied at t, if and only if (a) S asserted at t that p, (b) S actively believed at t that not p.

Falkenberg (1982: 135–136) proposed a distinction between strong and weak lies along the following lines. If S asserts that p while believing that not p, this is called a strong (Germ. ‘stark’) lie. However, when S asserts that p while neither believing p nor not p, then this counts as a weak (Germ. ‘schwach’) lie. For example, the speaker is unsure about what is the case, or they are ignorant in this respect, but they nevertheless assert that p. The problem is that standard definitions of assertion require that the speaker believes that p. But even when this condition is dropped, the Quality maxim requires that the speaker is truthful. Hence it may be asked whether a weak lie, as Falkenberg defines it, is an assertion at all. My guess is that it is rather the utterance of an assumption. If I do not know whether I failed a test or not, but nevertheless, when asked, say that I passed it, I utter the assumption that I passed the test. Only when I guess that I failed the test but say that I passed it, have I lied; this is a strong lie because I am inclined to believe that I failed the test: it is not the case that I neither believe p nor believe not p. We have already illustrated the different instances of being or not being truthful with the example of a speaker S, called Barbie, who utters I’m suffering from a heart disease (cf. Falkenberg 1982: 54–58, building on an example by E. Utitz) in section 3.7. This example will be repeated here for sake of convenience. First, Barbie believes at t that she is suffering from a heart disease and it is indeed the case that she is suffering from a heart disease. The propositional content p is true and the speaker is truthful (she does not lie). Second, Barbie believes at t that she is suffering from a heart disease but, actually, she is not. Thus, she is mistaken. The propositional content p is false, but the speaker is truthful (she does not lie). Third, Barbie does not believe at t that she is suffering from a heart disease, and she knows that, as a matter of fact, she is not suffering from a heart disease. It follows that she lies (she is not truthful). Fourth, Barbie does not believe at t that she is suffering from a heart disease, but in fact, she is. She is mistaken since the propositional content p is true. Moreover, she lies (she is not truthful) since she wants to deceive. As stressed before, it is very important that a theory of lying can deal with this last case. With regard to the speaker’s intention to deceive, it does not matter what actually is the case; it only matters what the speaker believes to be the case. As a consequence, definition (2) refers to (weak) ‘believing’ instead of (strong)

104   

   Defining lying

‘knowing’. The notion of ‘actively believing’ in definition (2) intends to exclude those cases in which speaker S is merely accidentally or mistakenly in a certain state of belief (Falkenberg 1982: 45–50). To give an example, consider the case of someone who asserts that they are a student while forgetting or not knowing that their registration was cancelled the day before.

4.3 Intention to deceive Let us shortly compare our definition of lying to the recent proposal of Carson (2010). Through a series of careful redefinitions on the basis of prototypical lying scenarios, he finally ends up with the following definition of lying (his “L7”): (3)

Lying (Carson 2010: 37) A person S tells a lie to another person S1 iff: 1. S makes a false statement X to S1, 2.

S believes that X is false or probably false (or, alternatively, S does not believe that X is true), and

3.

S intends to warrant the truth of X to S1.

Definition (3) is intended to implement the following two basic tenets of Carson’s approach: (i) lying does not require that the liar intends to deceive others; (ii) lying requires making a statement that one warrants to be true (Carson 2010: 15). Requirement (ii) is connected with the basic assumption that lying is essentially a breach of trust. Most speech-act theoreticians would wholeheartedly agree with this conviction. But warranting the truth of p is of course built into standard definitions of assertion (or statement, the difference between the two often being neglected), be it in the style of Searle (1969) (constitutive rule: “Counts as an undertaking to the effect that p represents an actual state of affairs”) or as (1b) above. Therefore, “warranting the truth” follows from a reasonable taxonomy of speech acts. As for (i), I agree that deception should not directly be written into definitions of lying; the deception follows from the lack of truthfulness on the part of the liar. In (2) above, we made no reference to the deceptive intention of the speaker. But this seems to follow from the definitions (1) and (2). If a speaker intentionally asserts p while they do not believe p, then this counts as deceptive behavior. The speaker is not truthful, although they should be truthful, at least according to the first maxim of Quality. Thus, we do not need to build something like “Speaker

Intention to deceive   

   105

wants to deceive the Hearer” into the definition of lying  – this automatically follows from the requirement of being truthful. However, is it correct that lying does not require that the liar intends to deceive others? To support his claim, Carson (2010: 20–23) invents three examples in which a person can lie without intending to deceive anyone. All of these are not really convincing to me. Let us focus on the first example which goes as follows: “Suppose that I witness a crime and clearly see that a particular individual committed the crime. Later, the same person is accused of the crime and, as a witness in court I am asked whether or not I saw the defendant commit the crime. I make the false statement that I did not see the defendant commit the crime, for fear of being harmed or killed by him. However, I do not intend that my false statements deceive anyone. (I hope that no one believes my testimony and that he is convicted in spite of it.)” (Carson 2010: 20)

Carson (2010: 20) goes on to comment: “Deceiving the jury is not a means to preserving my life. Giving false testimony is necessary to save my life, but deceiving others is not; the deception is merely an unintended ‘side effect’. I do not intend to deceive the jury in this case, but it seems clear that my false testimony would constitute a lie.” However, I assume that the witness has deceived the jury as well as the defendant – both are deceived because the witness asserts that s/he has a certain belief which s/he actually has not. Hinting at the possible fact that the witness does not actually hold that belief and that s/he only pretends to have that belief, is, according to my linguistic approach, not a “side effect”, and may be best analysed as a conversational implicature. But this concept is not available within Carson’s framework. Thus, there is no need to skip condition (1c) in my above definition. In a recent discussion of my argument, Dynel (2011: 151) comments that contrary to Carson’s claims, the unintended side effect “cannot count as a lie, thanks to the speaker’s communicative intention not to deceive the jury. Nor is it the case that successful but unintended deception of the jury can be perceived as an implicature, as Meibauer (2011) seems to suggest. This is because implicatures must be intended by the speaker […].” My view is indeed that an (of course intended) implicature of the type ‘S asserts p but wants me to think that I do not believe p’ is possible, perhaps induced by the maxim of Relevance. This ties in with the concept of ‘hinting at’ or ‘intimating’ (Germ. ‘etwas durchblicken lassen’) (cf. also the concept of ‘insinuating’ in Fricker 2012). Dynel (2011) refers to Saul (2012) who introduces the notion of ‘misleading’, a notion that she considers a notion of common sense that is well-distinguished from lying on the one hand, and accidental falsehood on the other. Consider the utterance (4):

106   

(4)

   Defining lying

There is no improper relationship.

Imagine that (4) is spoken by Clinton in the Lewinsky trial. Assume that Clinton did have a sexual relationship with Monica Lewinsky. Saul (2012: 3) points out: “We know for a fact that some would report this utterance with ‘Clinton said there was never an improper relationship’, since they did so. And this is indeed reasonable, using a loose understanding of ‘say’ and bearing in mind that the context was one in which a possible past relationship was being discussed.” However, for Saul (2012: 2), (4) is not a case of lying, but a case of (perhaps deliberate) misleading. In general, Saul thinks that if a speaker leads their audience into a false belief but says only things s/he knows to be true, then s/he has not lied. The basic definition of lying that is given in Saul (2012) is similar to the one given by Carson (2010) in that it includes the concept of ‘warranting’: (5)

Lying (Saul 2012: 19) If the speaker is not the victim of linguistic error/malapropism or using metaphor, hyperbole, or irony, then they lie iff (1) they say that P; (2) they believe P to be false; (3) they take themself to be in a warranting context.

A warranting context is one in which sincerity is expected (Saul 2012: 19). However, instead of assuming condition (5(3)), it makes more sense to derive sincerity from felicity conditions for speech acts or from the operation of the communicative principle and the conversational maxims. In addition, Saul (2012: 65) proposes a “complete” definition, i.e., one that integrates notorious cases of underdeterminacy, as discussed widely in the contextualist literature. I will discuss these issues in section 4.7. The assumption that cases like (4) are not cases of lying but cases of misleading will be challenged in section 4.5. As said before, I would like to argue that these utterances give rise to conversational implicatures and are thus part of the total signification of the utterance (Grice 1989b: 41). I will then propose an extended definition of lying that includes intentionally false implicature.

Bald-faced lies and self-deception   

   107

4.4 Bald-faced lies and self-deception There are two phenomena that could be put forward against the plausibility of the definition of lying in (2). The first is the case of bald-faced lies as described by Sorensen (2007), and the second is the case of self-deception. According to Sorensen, in bald-faced lies the speaker presents a proposition p as true, does not believe that p is true, yet does not deceive the hearer, for it is mutually known to the participants of the discourse that p is false. In self-deception, it is not the hearer who is duped but the speaker. But how it is possible to lie to oneself if the liar knows the truth, i.e., represents something they believe is false to themself in a way that they believe it is true?

4.4.1 Bald-faced lies As his introductory example, Sorensen (2007) presents a little dialogue between the official guide Takhlef and the Norwegian reporter Asne Seierstad which is situated in Iraq under the dictatorship of Saddam Hussein. In this exchange it is manifest that Takhlef is insincere: (6)

Takhlef: Everything [President Saddam Hussein] did in the past was good and everything he will do in the future is good. Seierstad: How can you be so sure about that? Takhlef: I know it as a result of my belief in the party and his leadership.

Takhlef, Sorensen (2007: 252) points out, “is not merely pretending to assert that Saddam’s leadership is perfect” since “he wants to be on the record” and “defends the proposition by words and deeds”. Yet his utterances are bald-faced lies because it is mutually known (or so assumes Seierstad) in the context that Takhlef does not believe the propositions of his utterances in (6). In order to account for cases like (6), Sorensen (2007) proposes a substitute for the concept of the intent to deceive, namely a distinction between narrow and wide plausibility. Assuming that “lying is just asserting what one does not believe” (Sorensen 2007: 266), he distinguishes between assertions with narrow plausibility, where “someone who only had access to the assertion might believe it” and wide plausibility, understood as “credibility relative to one’s total evidence” (Sorensen 2007: 267). Bald-faced lies are then a case of narrow plausibility, as in the assertions of a lawyer defending a guilty client in which neither the lawyer themself nor the jury is supposed to believe in the truth of these assertions.

108   

   Defining lying

However, the point with the lawyer’s assertions is that s/he describes possible scenarios in order to force the jury to provide evidence for something the client is supposed to have done. So Sorensen (2007: 255) is mistaken when he argues that “when the lawyer does not believe the conclusions he is lying”. Similarly, the utterances of Takhlef are not bald-faced lies but assertions since his propositions may be presented with narrow plausibility and are not, as Sorensen (2007) rightly observes, self-defeating: It is implausible to say Everything Saddam did is good but I do not believe it or the like. Hence, Seierstad may be wondering to which extent her guide is truthful. I conclude, then, that Takhlef’s utterances are either a case of assertion with narrow plausibility, a case of self-deception, or a case of a cynical assertion (Sorensen 2007: 260–261, drawing on Kenyon 2003). Here, the speaker is said to be “manifestly insincere”. I take it that his manifest insincerity is a conversational implicature. For instance, Takhlef wants Seierstad to understand that he is under pressure to assert that Saddam Hussein is a good leader, but in reality he does not believe this and even may be someone who opposes Saddam’s dictatorship. Carson (2010) defends his view that asserting amounts to warranting the truth of p against Sorensen. Both agree that the intent to deceive is not a necessary precondition for lying. However, warranting the truth is also in the scope of Sorensen’s approach to bald-faced lies since he defines lying simply as asserting the false. Carson (2010) takes up Sorensen’s example of an interchange in an Iraqi civilian hospital reported by Seierstad. The context is that Seierstad sees a ward with wounded soldiers, indicating that the military hospitals are already overcrowded. (7)

Seierstad: How many soldiers have you admitted today? Doctor: There are no soldiers here. Seierstad: But they are wearing uniforms? Doctor: I see no uniforms. You must go now, do you hear?

I consent to Carson (2010: 35) who argues that “the doctor still warrants the truth of what he says since he invites Seierstad to trust and rely on the truth of what he says and guarantees the truth of what he says.” If what he says is evidentially false, I would like to add that he infringes the maxim of Quality, thus inviting the pragmatic inference that he is not entitled to admit the over-crowdedness of the hospital. In the philosophy of language, several authors assume that so-called baldfaced lies and knowledge-lies are counterexamples to the idea that lying involves an intention to deceive (e.g., Carson 2010, Sorensen 2007, 2010, Fallis 2009; see also Stokke 2013a, b). Recently, Lackey (2013) has argued against the “unhappy

Bald-faced lies and self-deception   

   109

divorce” of lying and deceit/deception, yet she takes it for granted that bald-faced lies are genuine lies. While I side with her in that I think that lies are deceiving, I still argue that bald-faced lies are no lies. Note that Kenyon (2003) holds a similar view with respect to what he calls “cynical assertion”. This position is dismissed by Lackey (2013: 238, Fn. 7) largely for intuitive reasons. Lackey (2013) introduces a distinction between deception and deceit and argues that the case of the cheating student sketched by Carson (2010) is a case of concealment, i.e., a case of deception. In contrast, Fallis (2013) tries to show that the cheating student does not deceive. In Meibauer (2013c), bald-faced lies are not counted as lies because, by definition, it is mutually known by the participants that what the speaker says is false. To frankly demonstrate that one is opting out from the operation of the Cooperative Principle and the maxims amounts to a severe attack on the addressee’s face and is impolite to a high degree (Mooney 2008). Therefore, I propose that baldfaced lies are special kinds of insults (see Neu 2008, Mateo and Yus 2013).

4.4.2 Self-deception In our definition of lying, we implicitly assumed (via the addressee H mentioned in the definition of assertion) that lying presupposes a liar and someone who is lied to. But it may be asked whether it is possible that speaker A lies to themself. This is the riddle of self-deception that we will shortly touch upon. If self-deception is modelled on lying, we have to make a proper distinction between self-deception, and mere error or accidental falsehood. But in this model, we are forced to assume that “self-deceivers apparently must (1) hold contradictory beliefs, and (2) intentionally get themselves to hold a belief they know or believe truly to be false” (Deweese-Boyd 2006: 2). Thus, we are faced with two paradoxes: “One concerns the self-deceiver’s state of mind – the so-called ‘static’ paradox. How can a person simultaneously hold contradictory beliefs? The other concerns the process or dynamics of self-deception – the so-called ‘dynamic’ or ‘strategic’ paradox. How can a person intend to deceive herself without rendering her intentions ineffective?” (Deweese-Boyd 2006: 2) It seems impossible for one to believe, in full awareness, a proposition and its negation simultaneously. Thus, one may stay sceptical about the possibility of self-deception. However, when the dynamics of self-deception is considered, it is more plausible to assume a state in which a person believes that p and another state in which they believe that non-p. But there are problems here, too, as Deweese-Boyd (2006: 2) points out:

110   

   Defining lying

“As the deceiver, she must be aware she’s deploying a deceitful strategy; but, as the deceived, she must be unaware of this strategy for it to be effective. And yet it is difficult to see how the self-deceiver could fail to be aware of her intention to deceive. A strategy known to be deceitful, however, seems bound to fail. How could I be taken in by your efforts to get me to believe something false, if I know what you’re up to? But if it’s impossible to be taken in by a strategy one knows is deceitful, then, again, self-deception as it has traditionally been understood seems to be impossible as well.”

Attempts at the solution to these paradoxes are either intentionalist approaches or non-intentionalist approaches. Intentionalist approaches usually work with temporal or psychological partitioning. An example that illustrates how one can simply change beliefs over time (temporal partitioning) is provided in (8): (8)

An example of self-deception (Deweese-Boyd 2006: 3) “(…) an official involved in some illegal behavior might destroy any records of this behavior and create evidence that would cover it up (diary entries, emails and the like), knowing that she will likely forget having done these things over the next few months. When her activities are investigated a year later, she has forgotten her tampering efforts and based upon her falsified evidence comes to believe falsely that she was not involved in the illegal activities of which she is accused.”

Obviously, the static paradox does not arise here because the official never simultaneously holds contradictory beliefs. Forgetting or suppressing things, the normal degradation of memory, the tendency to believe what one practices – all this may play a role when one changes one’s mind about a certain event. But as far as I see, the dynamic paradox still seems to be there. Psychological partitioning “is the division of the self into psychological parts that play the role of the deceiver and deceived respectively” (Deweese-Boyd 2006: 4). Intentionalists who favor psychological partitioning tend to assume that it is possible to believe in p on the one hand, and to believe in non-p on the other hand, but, of course, not to believe in p & -p. Yet contradictory belief is accepted. This may solve, in a way, the static paradox. But it is still hard to explain the initial intention to substitute -p for p since this intention motivates the partitioning in the first place. Thus, the dynamic paradox seems to survive. The non-intentionalist approach basically consists in arguing that “to be deceived” is “nothing more than to believe falsely or be mistaken in believing” (Deweese-Boyd 2006: 5) – wishful thinking, the desire for things to be like one wishes them to be, is thus a close relative of self-deception. It appears that both paradoxes do not arise in this approach; however, the close parallel of self-deception to the interpersonal model of lying dissolves into thin air. In sum, then, there

Bald-faced lies and self-deception   

   111

are good reasons for being sceptical about the concept of self-deception if modelled on our definition of lying. In this connection, a quick glance at confabulation is also of interest. In his introductory chapter, Hirstein (2005: 1) gives a telling example of what confabulation looks like: (9)

A case of confabulation (Hirstein 2005: 1) A neurologist enters a hospital room and approaches an older man sitting up in bed. The neurologist greets him, examines his chart, and after a brief chat in which the man reports feeling fine, asks him what he did over the weekend. The man offers in response a long, coherent description of his going to a professional conference in New York City and planning a large research team, all of which the doctor writes down. The only problem with this narration is that the man has been in the hospital in the entire weekend, in fact for the past three months. What is curious is that the man is of sound mind, yet genuinely believes what he is saying. When the doctor informs him that he is mistaken, he replies “I will have to check with my wife about that,” then seems to lose interest in the conversation. The man isn’t “crazy” or schizophrenic; he is quite coherent and can answer all sorts of questions about who his children are, who the current president is, and so on. He is confabulatory, owing in this case to the fact that he has Korsakoff’s syndrome, a disorder that affects his memory, producing a dense amnesia for recent events. But unlike other patients with memory dysfunction, who freely admit that their memories are poor, a patient with Korsakoff’s syndrome will confidently report as memories events that either did not happen (or at least did not involve him) or that happened to him, but much earlier in life. This man’s act of describing the conference in New York City is known as confabulation.

The term ‘confabulation’ was first applied to Korsakoff’s patients by the German psychiatrist Karl Bonhoeffer in the early 1900s. The term comes from the Latin verb confabulari (literally ‘together + talk’). The German verb fabulieren means, according to Wahrig-Burfeind (2006, 8th edition), „Fabeln erdichten, Geschichten erfinden, fantasievoll u. wortgewandt erzählen“. In this paragraph, we will shortly go into two questions: First, what is he difference between lying and confabulation? Second, is self-deception confabulation (or vice versa)? In his careful attempt at providing a clear-cut definition of confabulation as an epistemic phenomenon, Hirstein tentatively proposes the following definition:

112   

(10)

   Defining lying

Confabulation (Hirstein 2005: 187) Jan confabulates if and only if 1. Jan claims that p. (e.g., Jan claims that her left arm is fine.) 2. 

Jan believes that p.

3. . Jan’s thought that p is ill-grounded 4. 

Jan does not know that her thought is ill-grounded.

5. 

Jan should know that her thought is ill-grounded.

6. 

Jan is confident that p.

Using “to claim” instead of, for instance, “to assert” in the first condition, is intended to cover non-verbal confabulations, e.g., acts of false pointing to pictures or showing something. The notion of “ill-groundedness” in the third, fourth, and fifth condition refers to the observation that it is not falsity in the strict sense that matters here – some confabulations may be (accidentally) true –, but “a malfunction in a knowledge-gathering process, either in perception or memory” (Hirstein 2005: 190). That Jan is “confident” that p reflects the fact that confabulators are – even when interrogated – quite sure that the story they told is true. It is obvious – and it follows from conditions 2. and 4. – that the confabulator has no intent to deceive. And this is the main difference from lying. Drawing on the work of Davidson (1986) and Mele (2001), Hirstein (2005: 221– 226) discusses whether self-deception might be understood as confabulation. He essentially views the difference between the two as a difference in a potentially greater access to the truth on the side of the self-deceiver: “In both confabulation and self-deception, the problem is that the warrant for the crucial thought is weak, for certain reasons. The self-deceived person’s belief is not based on the most reliable belief-forming processes his brain could supply if it were not in a self-deceptive mode. The clinical confabulator’s belief is not based on the most reliable belief-forming processes his brain could supply if it were not damaged. The primary difference between the two phenomena is that the self-deceived person often has a greater ability to access the information or processes that are needed to show that his belief is ill-grounded, whereas in the confabulator, the normal routes of access have been destroyed.” (Hirstein 2005: 213–14)

What the self-deceiver has and the confabulator does not have, is a certain tension between certainty and doubt with respect to the facts. Hirstein (2005: 215) proposes a scale of the degree of tension that comprises several disorders, with the neutral normal person as the one who is optimally balanced out.

Lying and falsely implicating   

(11)

   113

Scale of the degree of tension (certainty/doubt) (Hirstein 2005: 215) Clinical confabulator Sociopath Self-deceived normal person without tension Normal confabulator Neutral normal person Self-deceived normal person with tension Lying normal person Obsessive-compulsive normal person Clinical OCD sufferer

The clinical confabulator has a very low level of tension, and the Clinical OCD sufferer (OCD = obsessive-compulsive disorder) has a very high level of tension. Let us have a closer look on this taxonomy. A clinical confabulator, as opposed to the normal confabulator, is a person that suffers from orbitofrontal damage. Confabulation, and this is one of Hirstein’s most interesting findings with respect to lying, occasionally shows up even with normal people, for instance, “young children, subjects to hypnosis, people asked to justify opinions, and people reporting mental states” (p. 8). Sociopaths are people, possibly with some sort of orbitofrontal damage, who show incorrect social behavior and a lack of empathy. Some of them “are capable of lying skilfully, but they come close to confabulators in that they sometimes believe what they are saying.” (p. 92) Finally, persons with OCD are people with orbitofrontal hyperfunction (Hirstein 2005: 97–99). Their main property is pathological doubt. Doubt, for instance, whether a door is locked, whether the oven is turned off, etc. Such patients are forced to check things; the mere memory of an action is insufficient. It could be speculated that this strange behavior has to do with a hyperactive working memory process. In sum then, Hirstein (2005: 226) assumes that there is indeed a close relation between self-deception and confabulation since “when a self-deceived person who experiences no tension asserts a self-deceptive belief, he is confabulating”. I conclude, then, that neither bald-faced lies nor self-deception are counterarguments of my definition.

4.5 Lying and falsely implicating In this section, I am going to argue that it is possible to lie while saying the truth. This argument is possible, and indeed necessary, when one considers the fact

114   

   Defining lying

that many assertions are connected with conversational implicatures. These implicatures are intended by the speaker and are intended to be derived by the hearer. If the implicatures are believed by the hearer, then it is of course possible to lie by intentionally using false implicatures. Thus, we get an extended understanding of lying that includes assertions as well as implicatures. Before we will go into the proper definition of conversational implicature, we will shortly discuss two questions that are intimately related, and directly bear on the notion of conversational implicature. The first question is how normative conversational implicatures are; the second is whether all conversational implicatures are cancellable. Both questions are heavily disputed in recent literature (Saul 2002, Weiner 2006, Davis 2007, Blome-Tillmann 2008, Jaszczolt 2009, Burton-Roberts 2010, 2013).

4.5.1 Normativity and cancellability In order to point out what the connection between these two questions is, we will first have a look at the classical characterization of the notion of conversational implicature in Grice (1989a: 30–31): (12)

Conversational implicature Grice (1989a: 30–31): “A man who, by (in, when) saying (or making as if to say) that p has implicated that q, may be said to have conversationally implicated that q, provided that (1) he is to be presumed to be observing the conversational maxims, or at least the Cooperative Principle; (2) the supposition that he is aware that, or thinks that, q is required in order to make his saying or making as if to say p (or doing so in those terms) consistent with this presumption; and (3) the speaker thinks (and would expect the hearer to think that the speaker thinks) that it is within the competence of the hearer to work out, or grasp intuitively, that the supposition in (2) is required.”

Davis (2007: 1658) dubs the first condition the ‘cooperative presumption condition’, the second condition the ‘determinacy condition’, and the third condition the ‘mutual knowledge condition’. What is of interest here is the determinacy condition. We may ask what the status of that condition is. Take a standard case like Ken’s utterance I’ve seen Mr Smith with a woman tonight, implicating ‘The woman was not his wife, sister, mother, nor another close relative, etc.’ Then we attribute to Ken the intention to convey this implicature. After all, if he had not intended us to draw this conclusion, he should have

Lying and falsely implicating   

   115

said I’ve seen Mr Smith with his wife tonight. Thus, the implicature is determined in a certain way, as the determinacy condition requires. What, then, is the normative aspect of a conversational implicature? Davis’ proposal is to define the notion of ‘proper implicature’ (Davis 2007: 1662): (13)

Proper implicature (Davis 2007) S properly implicates that p when S not only means or implies that p by saying something else, but does so properly, the way one should do it, thereby fulfilling one’s communicative responsibilities.

Note that Davis (2007) has a very broad view of implicature in which implicature is similar to indirection – meaning or implying p when saying something else. Saul (2002) attacks this very broad view by presenting absurd scenarios, e.g., when Bettina presents an aspirin to Albert in order to convey the message that he should flee because aliens are in sight. These kinds of examples are excluded when the notion of properly implicating is taken into account. Nevertheless, Davis insists that conversational implicatures arise because of their conventional nature in the first place. Saul (2002), in contrast, claims that speakers make their implicatures available to the public. Davis certainly has a point in remarking that even in standard cases of conversational implicature there is a range of potential implicatures that is far from being determinated. For instance, if a National Transportation Safety Board official, when arriving at a crash scene, utters Some passengers died, several implicatures are compatible with this particular statement; they could have been implicating the negations of all the statements under (14): (14)

a.

All but one of the passengers died.

b.

A few passengers died.

c.

90% of the passengers died.

d.

All passengers died instantly.

e.

All passengers, crew members, and animals on board died.

Davis (2007: 1665) concludes: “The list could easily be extended. In each case, it is no less reasonable to reason that if the official knew it to be true, he should have said so; ergo he must have meant that it is not true. But the official is not likely to have implicated the negations of any of these statements. Grice’s theory fails to differentiate between actual implicatures and countless things speakers might have implicated but did not. The theory ‘overgenerates’ implicatures if it gener-

116   

   Defining lying

ates any. In fact, though, the possibility of some of the implicatures shows that the determinacy condition is rarely satisfied.” While I agree that the negations of (14) are candidates for implicature in this case, I do not see why the determinacy condition must be altogether wrong. After all, utterances have an implicature potential (similarly to the illocutionary potential of an utterance), and the best candidate must be singled out via the operation of the working out schema: (15)

Working out schema (Grice 1989a: 31) 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. To work out that a particular conversational implicature is present, the hearer will rely on the following data: (1) the conventional meaning of the words used, together with the identity of any references that may be involved; (2) the Cooperative Principle and its maxims; (3) the context, linguistic or otherwise, of the utterance; (4) other items of background knowledge; and (5) the fact (or supposed fact) that all relevant items falling under the previous headings are available to both participants and both participants know or assume this to be the case. A general pattern for the working out of a conversational implicature might be given as follows: “He has said that q; there is no reason to suppose that he is not observing the maxims, or at least the Cooperative Principle; he could not be doing this unless he thought that p; he knows (and knows that I know that he knows) that I can see that the supposition that he thinks that p is required; he has done nothing to stop me thinking that p; he intends me to think, or is at least willing to allow me to think, that p; and so he has implicated that p.”

For Davis (2007: 1667), however, a typical scalar implicature like Some … +> ‘Not all …’ is nothing but a matter of convention: There is a “regular use” “that is socially useful, self-perpetuating, and to some extent arbitrary”. The convention exists because people have “a habit of doing so” and “people before them have done so”. Admittedly, there may be a conventional aspect of the otherwise conversational implicature, an aspect that is also meant when one speaks of the generalized or default character of implicatures; but when it comes to explaining the determinacy condition, the appeal to mere conventionality appears quite helpless. But this is, in fact, the credo of Davis (2007: 1668): “The fact that a conversational implicature is conventional suffices to account for its propriety. Satisfaction of the determinacy condition is unnecessary.”

Lying and falsely implicating   

   117

For Saul (2002), there may be mismatches between what the speaker intended to convey on the one hand, and what the addressees took up on the other hand. This shows that the determinacy condition is indeed problematic. But for her, it is not conventionality that matters – as Davis would have it –, but availability. Besides, note that Grice already spoke of the condition “that all relevant items falling under the previous headings are available to both participants” [my emphasis, JM]. The reason for Saul’s assumption that “making available of an implicature” is the notion that matters, is twofold: On the one hand, there are intended implicatures that are not yet derived, and on the other hand, there are unintended implicatures that are nevertheless derived. For Saul, this amounts to postulating two new categories: (a) ‘utterer-implicatures’ which fail to be conversational implicatures, and (b) ‘audience-implicatures’ which fail to be utterer-implicatures. Here are two examples for each type, modelled after Grice’s famous letterof-recommendation case: Utterer implicature. S wants to implicate q, but fails to do so. Utterer-implicatures are intended implicatures; however, for one reason or another, the implicature is not derived by the audience. Here is an example: (16)

[Fred] “I have been asked to write a letter for my student, Fred, who is a poor philosopher and a chronic petty thief. I don’t want to put such things in writing, but I want to write truths and to communicate that Fred should not be hired. Since Fred does happen to be a good typist, I write only this, trusting Fred’s prospective employers (who I take to be a philosophy search committee) will discern my message. Unbeknownst to me, Fred is applying for a job as a typist. What I wrote was precisely the information that Fred’s prospective employers require. They don’t need to assume that I think anything other than what my words mean in order to preserve the assumption that I am being cooperative. What I mean to communicate, then, fails to be implicated.”

Audience implicature. S does not want to implicate q, but the audience takes him to do so. Hence, “audience-implicatures are just like conversational implicatures except that audiences have authority over what is audience-implicated.” (p. 10) Here is an example:

118   

(17)

   Defining lying

[Felix] “Imagine, for example, that I am writing a letter of reference for my student Felix, who I consider to be an excellent philosopher. I falsely believe, however, that he is applying for a job as a typist, so I discuss nothing but his fine typing and punctuality. The audience, which is actually a philosopher appointments committee, takes me to have conversationally implicated that Felix is a poor philosopher. They are, of course, wrong: clause (3) of Grice’s characterisation [see (12)] was not satisfied, as I had no idea that they would, or even could, work out from my utterance that I think Felix is a poor philosopher, and I would not have made my utterance if I’d realised the situation. This claim, then, fails to be conversationally implicated.”

What Saul convincingly shows with these and further examples is that all sorts of misunderstandings may happen when it comes to the attempt at conversational implicating something. All that the speaker can do is to try to make available a certain implicature. Because of the distinction between utterer implicature and audience implicature, the notion of a conversational implicature (which is not the mere intersection of utterer&audience implicature) gains still more interest: “There are, then, cases in which we can reasonably say that the audience should have worked out the conversational implicature, even if they failed to do so. This means that conversational implicature is a more normative notion than utterer&audience implicature.” (p. 11) A last example illustrates the point: (18)

[Wesley] “To see the interest of a notion like this one, imagine that I write a letter of reference for Wesley, another of my (fictional) unpromising students. I know that Wesley is applying for a philosophy job, and I write a letter designed to communicate my low opinion of Wesley. I write (truthfully), ‘Wesley’s main virtues as a philosopher are punctuality, an attractive choice of fonts, and an encyclopaedic knowledge of illegal pharmaceuticals.’ The audience, though certainly capable of working out from this that I think Wesley is a poor philosopher, reads too quickly, and takes away from the letter only the information that Wesley has encyclopaedic knowledge. They hire him, become disappointed, and complain to me.”

With respect to this scenario, Saul argues that the speaker cannot defend themself by saying that they utterer&audience-implicated the claim; but what they can do is to “maintain that I conversationally implicated it: It was required in order to understand me as cooperative, and my audience was capable of working this out.” (p. 12)

Lying and falsely implicating   

   119

The normative dimension of conversational implicature is thus to take responsibility for making available the information that is needed for deriving a certain implicature. Davis (2007: 1668) agrees with Saul “that if the determinacy condition were satisfied, the speaker would have made his or her implicature available, making it a proper implicature.” But this has, according to him, nothing to do with the soundness of the determinacy condition. What really matters is the existence of an implicature convention. I still find the gist of the determinacy condition quite convincing: The speakers try to convey a certain implicature while the hearers put forward a hypothesis about what the speakers tries to convey. Despite these cooperative efforts on both parts, misunderstandings may happen, as Saul’s examples impressively illustrate. However, these potential misunderstandings are constrained by “normative” aspects: Not every weird implicature is allowed, and the assumption that utterers usually only make reasonable implicatures available is in order. It goes without saying that this might have something to do with conventionality, but I think that the assumption that the utterer implicature is intended is still necessary. Conversational implicatures are intended by the speaker. Yes, they may be misunderstood, but more importantly for the definition of conversational implicature, they may be cancelled. A standard example is I have seen Mr Jones with a woman (+> ‘not his wife’, etc.) tonight; actually, I believe it is his wife (cancels ‘not his wife’). As is well-known, the criterion of cancellability is a core criterion for the status as conversational implicature since conventional meanings (including so-called conventional implicatures) are never cancellable. Hence, cancellability is a property that is very important for the task of correctly drawing the semantics/pragmatics boundary. In a recent article, Weiner (2006) argues that there are cases of implicature in which the cancellability test fails. Here is an example provided by Weiner (2006: 128) that is intended to prove the point: (19)

[Crowded train] “Suppose that Alice and Sarah are in a crowded train; Alice, who is obviously able-bodied, is sprawled across two seats, and Sarah is standing. Sarah says to Alice, ‘I’m curious as to whether it would be physically possible for you to make room for someone else to sit down.’ The implicature is that Alice should make room. It is extraordinarily unlikely that Sarah really is curious about whether Alice is physically capable of moving, since it is mutually obvious that she is capable.”

Thus, as Weiner points out, Sarah has flouted Grice’s first maxim of Quality. But the implicature may not be cancelled: “Suppose now that Sarah adds, ‘Not that you should make room; I’m just curious.’ This has the form of an explicit cancella-

120   

   Defining lying

tion of the implicature. Nevertheless, the implicature is not cancelled. Sarah is still suggesting, even more rudely, that Alice should make room.” And with her second remark, she flouts the first maxim of Quality again. One could also argue that, by her second remark, Sarah reinforces the implicature instead of cancelling it. I find examples like these quite convincing: In very many cases explicit cancelling “by the addition of a clause that states or implies that the speaker has opted out” (Grice 1989a: 39), e.g., by the addition of statements like but not p or I do not mean to imply that p, is, to say the least, not quite plausible. That the speaker was in fact ironical or metaphorical is mutually known to both participants because the speaker has made clear that this was their intention. But, of course, there is the possibility of cancelling through an imaginable alternative context in which the implicature does not arise in the first place. Thus, contextual cancelling is given “if the form of utterance that usually carries it is used in a context that makes clear that the speaker is opting out” (Grice 1989a: 39). Blome-Tillmann (2008: 159) accepts these difficulties with contextual cancelling. But for him, it is not an argument against cancellability altogether, because explicit and contextual cancelling are related. For every context in which explicit cancelling does not work, there is a possible contextual alternative in which it does work. Jaszczolt (2009) attacks this proposal in turn. According to her, explicit and contextual cancellation belong to different levels of language description that should be kept apart, hence “should not be discussed as part of the same test” (p. 262). Directed against Weiner (2006) and Blome-Tillmann (2008), she argues that “a particular scenario on which an implicature cannot be cancelled cannot ever constitute a counterexample to cancellability tout court” (p. 263–264). Moreover, Jaszczolt worries about the fact that explicit cancellability is a problematic criterion for those working with authentic discourse, and that it is hard to separate it from repairs (see also Mayol and Castroviejo 2013). And finally, she points out that Grice “himself did not regard cancellability as a decisive criterion in distinguishing implicatures” (p. 267), for example, when “using a word or form of words in a loose or relaxed way” (Grice 1989a: 44). I conclude from this discussion that there are indeed cases in which explicit cancelling is not possible. But this appears to have no repercussion on the general notion of implicature in which at least contextual cancelling should be possible.

4.5.2 A definition of conversational implicature According to the by now classical typology of implicatures developed by Grice (1989a), the total signification of an utterance consists of ‘what is said’ and ‘what is implicated’, the latter consisting of conventional implicatures and conversa-

Lying and falsely implicating   

   121

tional implicatures, with the conversational implicatures consisting of generalized and particularized implicatures, in turn (see Grice 1989b: 41). We do not need to deal with conventional implicatures here since they are bound to the literal meaning of expressions and therefore not cancellable. Hence, I will assume that conventional implicatures are the usual cases of lying or being truthful. For the sake of the following discussion, let us define a conversational implicature as in (20): (20)

Conversational Implicature S conversationally implicated at t that q, if and only if (a) S asserted at t that p, (b) S presented q as true, (c) q is calculable from the assertion of p, (d) q is cancellable.

Contrary to what is stated in (20a), asserting is not a precondition for conversationally implicating because implicating is also possible with other types of speech acts; e.g., the so-called existential implicature (formerly existential presupposition), as in Who comes? +> ‘Someone comes’, may be realized through wh-questions (cf. Meibauer 1991). However, since the standard examples in the literature are indeed assertions, the definition in (20) may be justified for the sake of the argument. In addition, it may be disputed whether there is asserting involved in the case of figurative utterances. I am inclined to assume that this is indeed the case because assertion is a precondition for deriving a conversationally implicated figurative meaning. Conditions (20c) and (20d) include the two most important criteria for conversational implicatures. Condition (20c) abbreviates “calculable with the help of the cooperative principle and the maxims”, and condition (20d) says that implicatures can be defeated by the addition of premises. One has to bear in mind that the calculation of a conversational implicature is actually a hypothesis. In the following discussion, I will assume that the hypothesis with regard to the conversational implicature derived by the hearer is identical to the conversational implicature intended by the speaker. Conversational implicatures, being no speech acts, do not have sincerity (or success) conditions. Whether they have truth conditions of their own seems open to dispute; in a recent review, Horn (2004: 25, Fn.1) leaves it open if this is the case. However, though q is cancellable by definition, it must be the case that the speaker who is implicating q presents q as true. Hence condition (20b).

122   

   Defining lying

In (20), we did not make a distinction between a Generalized and a Particularized Conversational Implicature. This distinction is illustrated in (21), an example taken from Levinson (2000: 16–17): (21)

Generalized Conversational Implicature (GCI) vs. Particularized Conversational Implicature (PCI) Context 1 A: What time is it? B: Some of the guests are already leaving. PCI: ‘It must be late.’ GCI: ‘Not all of the guests are already leaving.’ Context 2

A: B: PCI: GCI:

Where’s John? Some of the guests are already leaving. ‘Perhaps John has already left.’ ‘Not all of the guests are already leaving.’

Because the scalar implicature triggered by some arises in both contexts, it is relatively context-independent. Relative context-independence is the most prominent property of GCIs. In addition, GCIs are normally, or even consistently, associated with certain linguistic forms. In contrast to GCIs, PCIs are highly context-dependent, and they are not consistently associated with any linguistic form. It could very well be that these types of conversational implicature behave differently with regard to falsely implicating. Note that Burton-Roberts (2010: 147) argues that PCIs are never cancellable, while GCIs are. He draws a distinction between “implicatures that can arise independently of any intention of the speaker (GCI)” on the one hand, and “implicatures that arise only in virtue of the speaker’s intention to implicate (PCI)” on the other hand. The distinction is suggestive, but I refrain from splitting up the general definition of implicature according to this distinction. Burton-Robert’s main argument is that cancellation – understood as “cancellation without contradiction of intention” – is impossible for PCIs (Burton-Roberts 2010: 147): “Both GCIs and PCIs are context-dependent in the sense that they are – or are supposed to be – cancellable in, and by, the context of utterance. The crucial difference between them lies in the fact that GCIs are context-dependent only in that (negative) respect. PCIs, by contrast, are context-dependent in the further (positive) respect that they depend on a particular context to arise in the first place. In other words: with GCIs, context plays only a destructive (filtering, cancelling) role whereas, with PCIs, the particular context plays a constructive role. That is what makes them ‘particularised’. Outside any context of utterance, PCIs simply do not arise. By contrast, even when – in fact, especially when – an utterance gives rise to a GCI is de-contextualised, the GCI does still ‘arise’, in some intuitive sense.”

Lying and falsely implicating   

   123

Keeping this distinction in mind, we will analyze possible differences between falsely implicating with GCIs and PCIs.

4.5.3 “Lying while saying the truth”: Extended definition of lying Let us now concentrate on the relation between a lie and an implicature. Their interrelation is nicely shown in the following Story of the Mate and the Captain in (22) (cf. Posner 1980): (22)

The Story of the Mate and the Captain A captain and his mate have a long-term quarrel. The mate drinks more rum than is good for him, and the captain is determined not to tolerate this behaviour any longer. When the mate is drunk again, the captain writes into the logbook: Today, October 11th, the mate is drunk. When the mate reads this entry during his next watch, he is first getting angry, then, after a short moment of reflection, he writes into the logbook: Today, October 14th, the captain is not drunk.

The point is that the logbook entry of the mate is true. It is indeed the case that the captain is not drunk. However, a reader will understand that this is an exception because the captain is usually drunk most of the time. The calculation of the implicature starts from assuming a presumptive violation of the maxim of Relevance, for entries in logbooks must be relevant. In this story, we have a classical case of ‘lying while saying the truth’. What is important for the following discussion is the assumption that implicatures are additional propositions arising from the situation of the utterance. As defined in (20), the special properties of such propositions are cancellability and calculability. If conversational implicatures are additional propositions, these propositions should be either true or false. Indeed, this is what Grice (1989a: 39) observes, as becomes clear from the quote in (23): (23)

“Since the truth of a conversational implicature is not required by the truth of what is said (what is said may be true – what is implicated may be false), the implicature is not carried by what is said, but only by the saying of what is said, or by ‘putting it that way’.”

Assuming that the content of the assertion as well as the content of the conversational implicature may be true or false, we must consider the four pairs in (24):

124   

(24)

   Defining lying

a.

p is true and q is true

b.

p is true and q is false

c.

p is false and q is true

d.

p is false and q is false

The relevant propositions p and q may be paraphrased as in (25): (25)

p = ‘Today, October 14th, the captain is not drunk.’ q = ‘The captain is usually drunk.’

I proceed from the assumption that by uttering p, S intends that H derives q. Then, the implicature is an integral part of the total signification of the utterance, as Grice (1989b: 41) puts it. With regard to the following discussion of the Story of the Mate and the Captain, I restrict the notion of lie/truthfulness to the utterance of p, while restricting the notion of deception/non-deception to the implicature q. Note that this serves the ease of exposition only. In fact, I will argue that the notion of deception is superfluous (with respect to untruthful implicatures) when substituted by an extended definition of lying. In the first case, p is true and q is true. The captain is not drunk on October 14th, and it is true that he is usually drunk. In this case, the mate is not lying and he is not deceptive, at least in the sense that the implicature is not false. Note, however, that the mate misrepresents himself as being in a proper position to make the entry in the log, i.e., the coming about of the true implicature may be considered deceptive, too. In the second case, p is true and q is false. The captain is not drunk on October 14th, and it is false that the captain is usually drunk. The mate does not lie with regard to p, but he is deceptive with regard to q. As I will argue later, it seems possible to include this case into an extended definition of lying. The third case, in which p is false and q is true, is more complicated. The captain is indeed drunk on October 14th. On the one hand, the mate may be mistaken in this regard (if he had checked the captain’s cabin, he would have seen that the captain was completely drunk.) If the mate is mistaken, he is truthful with regard to p. On the other hand, it may be the case that the mate is not truthful with regard to p. He definitely knows that the captain is drunk on October 14th, but he suppresses this knowledge in the hope that his implicature will be calculated more smoothly by the reader. Thus, he lies with regard to p in order to convey the implicature which is true in his eyes.

Lying and falsely implicating   

   125

The same distinction holds for the fourth case, in which p is false and q is false. With regard to p, the mate may be mistaken (then he is truthful) or not (then he is lying). With regard to q, however, the mate is deceptive. Let us now consider the relation between p and q in more detail. It is clear that the implicature q is cancellable by definition. The notion of cancellability is twofold: A conversational implicature may be explicitly cancelled, i.e., “by the addition of a clause that states or implies that the speaker has opted out”, or it may be contextually cancelled, i.e., in a different context the implicature may not arise (Grice 1989a: 39). The possibility of explicitly cancelling the implicature may be a motive for the speaker to prefer lying while saying the truth rather than simply lying because the former gives them the chance of withdrawing in case of detection. However, in utterance situations in which it is mutually agreed upon that a certain implicature has been conveyed, an attempt at cancelling the implicature is not convincing, to say the best. The logbook entry simply does not make sense if not associated with a conversational implicature. Look at (26): (26)

Today, October 14th, the captain is not drunk. This is not to say that he is usually drunk.

It is obvious that such a sequence gives rise to further questions: What other things are alluded to? Why does the mate not say the things he wants to say in a straightforward manner? Because q is intimately connected to p and the calculation of q is evidently intended by the speaker, the cancellation of q is not plausible in many situations. To be sure, more colloquial cancelling clauses may work better in this respect (e.g., Actually, he is never drunk), but this simply shows that there are degrees of cancellability that are dependent on the situation of an utterance. In order to capture the case of lying while saying the truth, it seems reasonable to propose the following extended definition of lying. Note that this definition makes no use of the notion of ‘deception’. (27)

Lie – extended definition (IMPL) S lied at t by uttering the declarative sentence σ iff (a) the definition of the lie holds, (b) or S thereby conversationally implicated that q, but actively believed that not q.

126   

   Defining lying

We do not have to restrict the definition in (27) to those cases in which explicit cancelling seems dubious because contextual cancelling works anyway and is indeed necessary for the general definition of conversational implicature. Falkenberg (1982: 137) tentatively proposes the following definition of ‘indirect lying’: “A lied indirectly, iff (a) A asserted that p and thereby implicated that q, (b) A believed that not q.” [my translation, JM]

Falkenberg rejects this definition again because he conceives it as being inconsistent with the logical principles of double negation, bivalency, and the excluded third. Maybe he has cases of irony in mind, where q equals -p. As I want to show in this chapter, this should not be considered as a severe problem as long as the distinction between assertion and implicature is upheld. What matters is that the parallel with indirect speech acts is not fully convincing because the proposition q is not a primary illocutionary act. Furthermore, Falkenberg’s definition does not account for the fact that the assertion of p may be a lie, too. What (27) does is to try to capture the intimate connection between the assertion and the false implicature. Thus, false implicatures are not only bound to assertions, they are also “directly intended”. Adler (1997: 446) stresses this point when he states that “in falsely implicating, rather than lying, the outcome is still directly intended, not merely a foreseen consequence.” With the help of (27), we may summarize our analysis of the Story of the Mate and the Captain in (28). Note that the distinction between mistake/no mistake would apply to the cases (28a) and (28b), too, but this would not change the status as a lie or truthful utterance. (28)

‘Lie’ in the Story of the Mate and the Captain

Case

p

q

lie

(a) (b) (c)

true true false;

(d)

false;

true false true true false false

no yesq 1. 2. 1. 2.

1. 2. 1. 2.

mistake no mistake mistake no mistake

no yesp yesq yesp,q

Two further comments shall suffice here: First, note that the mate may be mistaken with regard to q, too. The mate wants to deceive the reader by intending them to derive q. The mate does not believe that q is true, but, as a matter of fact,

Lying and falsely implicating   

   127

it is true (as turned out during a later trial). Now, with regard to our extended definition, we would say that the mate is lying. Thus, there is a striking parallel between genuine lies (remember the case of the woman who mistakenly believes that she is not suffering from a heart disease) and lying by falsely implicating. Second, note that there is an interesting difference in the strength of the lie. In (28b), we have the case of “falsely implicating while saying the truth”. In (28d – no mistake) we have the case of “falsely implicating while saying the falseness”. It may be argued here that the behavior of the mate in the second case is shrewder. After all, he asserts that the captain is not drunk (contrary to his better knowledge that he is drunk indeed, which corresponds to his false implicature); if it turned out later that the captain was drunk, the speaker could say that he acted for the captains’s good; thus, he calculates that it is better not to tell the truth in order to convey the false implicature smoothly. Hence, a detailed analysis of the relation between p and q may uncover the relative strength (corresponding to the effort of the speaker) of lying.

Falsely implicating and the distinction between GCIs and PCIs In the Story of the Mate and the Captain we have the case of a Particularized Conversational Implicature (PCI) connected with the maxim of Relevance. Let us now ask whether Generalized Conversational Implicatures (GCIs) and Particularized Conversational Implicatures (PCIs) behave differently. We start by discussing some of the cases Levinson’s theory of GCIs deals with, and then proceed with some cases that Levinson either classifies as PCIs or leaves out altogether. Levinson develops his revision of Grice’s maxims from three heuristics that follow from the need to overcome the “fundamental bottleneck in the efficiency of human communication occasioned no doubt by absolute physiological constraints on the articulators” (Levinson 2000: 28). These heuristics are the following: (29)

Levinson’s heuristics a. Heuristics 1:

What isn’t said, isn’t.

b.

Heuristics

2:

What is simply described, is stereotypically exemplified.

c.

Heuristics

3:

What’s said in an abnormal way, isn’t normal; or Marked Message indicates marked situation.

128   

   Defining lying

Heuristics 1 corresponds to Levinson’s Q-Principle (see Quantity 1 in Grice’s framework), Heuristics 2 to Levinson’s I-principle (Grice’s Quantity 2), and Heuristics 3 to Levinson’s M-principle (Grice’s Manner 1 and 3). These three principles are said to trigger Generalized Conversational Implicatures. For these correspondences, and a typical example, see the following table. (30)

GCI deriving principles

Heuristics

Principles

Grice’s maxims

Example

Heuristics 1

Q-Principle

Quantity, 1

Q-Implicature: (a) Some colleagues were drunk. +> ‘Not all of them were drunk.’ (scalar implicature) (b) The doctor believes that the patient will not recover. +> ‘The doctor may or may not know that the patient will not recover.’ (clausal implicature)

Heuristics 2

I-Principle

Quantity, 2

I-Implicature: Anna turned the switch and the motor started. +> ‘Anna turned the switch and then/therefore the motor started.’ (conjunction buttressing)

Heuristics 3

M-Principle

Manner, 1 and 3

M-Implicature: Bill caused the car to stop. +> ‘He did this indirectly, not in the normal way, e.g., by use of the emergency brake.’ (periphrasis)

When inconsistent implicatures arise, they are “systematically resolved by an ordered set of priorities”, among them Q > M > I (Levinson 2000: 39). The Q-Principle is defined as follows (Levinson 2000: 76; see also Geurts 2010): (31)

Q-principle Speaker’s maxim: Do not provide a statement that is informationally weaker than your knowledge of the world allows, unless providing an informationally stronger statement would contravene the I-principle. Specifically, select the informationally strongest paradigmatic alternate that is consistent with the facts. Recipient’s corollary: Take it that the speaker made the strongest statement consistent with what he knows, and therefore that:

Lying and falsely implicating   

   129

a.

if the speaker asserts A(W), where A is a sentence frame and W an informationally weaker expression than S, and the contrastive expressions form a Horn scale (in the prototype case such that A(S) entails A(W), then one can infer that the speaker knows that the stronger statement A(S) (with S substituted for W) would be false […]

b.

if the speaker asserted A(W) and A(W) fails to entail an embedded sentence Q, which a stronger statement A(S) would entail, and {S, W} form a contrast set, then one can infer that the speaker does not know whether Q obtains or not (i.e., ~K(Q) or equally, {P(Q), P ~ (Q)} read as ‘it is epistemically possible that Q and epistemically possible that not-Q’)

The I-Principle mentioned in the Speaker’s maxim requires that a speaker should not be more informative than necessary (see below). Wherever possible, the speaker should build on stereotypical assumptions. In the Recipient’s corollary two cases are distinguished, namely scalar implicature involving Horn scales and clausal implicature involving contrast sets. Let us start with scalar implicatures. Given a scale with p as an informationally weak and q as an informationally strong element, the assertion of p implicates the negation of q. In such cases, the speaker is supposed to be as informative as possible, thus observing the Q-Principle (or the first maxim of Quantity). Therefore, they could not say more than they actually did and this means that the stronger statement does not hold. A classical example is the utterance p = Some colleagues were drunk implicating q = ‘Not all of them were drunk.’ If the speaker believes that all colleagues were drunk, then they could said to be lying. This is in line with definition (27). Such cases have correctly been described as ‘half-truths’ (Vincent and Castelfranchi 1981: 762), or as cases in which one says the truth but not the whole truth (Falkenberg 1982: 137). The speaker pretends to observe the first maxim of Quantity and expects the hearer to derive the relevant implicature; in fact, however, they suppress information and consequently violate the maxim. The revelation of only part of the truth follows from the fact that, in a prototypical scale, the stronger item entails the weaker item, e.g., if all colleagues were drunk, then some were drunk. Let us now check how scalar implicatures behave in the four situations sketched above. For ease of exposition, the case in which p is mistakenly asserted as false is omitted. We take the scale as an example, with p = Some colleagues were drunk, and q = ‘Not all of them were drunk’.

130   

   Defining lying

(a) p is true and q is true: The speaker asserts that some colleagues were drunk. S/he thereby conversationally implicates that not all of them were drunk. Both propositions are true, therefore no lie is involved. (b) p is true and q is false: The speaker asserts that some colleagues were drunk. This assertion is true. At the same time, s/he conversationally implicates that not all of them were drunk. As a matter of fact, they all were drunk, and the speaker knows that this was the case. Thus, s/he reveals only part of the truth, and is therefore lying. (c) p is false and q is true: The speaker asserts that some colleagues were drunk. In fact, all colleagues were not drunk (they were sober), and the speaker knows that. So there is a lie in relation to the false assertion p. By her/his very utterance, s/he conversationally implicates that not all of the colleagues were drunk. However, if this is true, this yields a contradiction because it is logically not possible to assume (on the basis of the true implicature q) that not all of the colleagues were drunk, and (on the basis of the false assertion p) that all colleagues were not drunk (sober). Therefore, the case of p being false and q being true is ruled out. (d) p is false and q is false: The speaker asserts that some colleagues were drunk. In fact, all colleagues were not drunk (they were sober), and the speaker knows this. Therefore, her/his utterance is a lie. By her/his very utterance, s/he conversationally implicates that not all of the colleagues were drunk. This implicature is false, too, for it is true that all of the colleagues were not drunk (sober). Consequently, this case counts as a lie. Now we focus on clausal implicatures. Let {know, believe} be a contrast set. Then p = The doctor believes that the patient will not recover implicates q = ‘The doctor may or may not know that the patient will not recover.’ (Levinson 2000: 110). Again, we check the four diagnostic situations. (a) p is true and q is true: If both p and q are true, there is no lie involved. (b) p is true and q is false: It is true that the doctor believes that the patient will not recover, but it is false that the doctor may or may not know that the patient will not recover. In fact, the speaker knows that the doctor knows that the patient will not recover, i.e., there is no epistemic uncertainty involved. Therefore, the speaker is lying. (c) p is false and q is true: It is false that the doctor believes that the patient will not recover. In fact, the doctor does not believe that the patient will not recover. Then the implicature that the doctor may or may not know that the patient will not recover cannot be true, i.e., there is no epistemic uncertainty involved. Thus, this case is ruled out.

Lying and falsely implicating   

   131

(d) p is false and q is false: It is false that the doctor believes that the patient will not recover. Moreover, it is false that the doctor may or may not know that the patient will not recover. Under both conditions it is true that s/he knows that the patient will not recover. Therefore, the speaker is lying. Now consider the I-principle (Levinson 2000: 114–115): (32)

I-Principle Speaker’s maxim: the maxim of Minimization. “Say as little as necessary”; that is, produce the minimal linguistic information sufficient to achieve your communicational ends (bearing Q in mind). Recipient’s corollary: the Enrichment Rule. Amplify the informational content of the speaker’s utterance, by finding the most specific interpretation, up to what you judge to be the speaker’s m-intended point, unless the speaker has broken the maxim of Minimization by using a marked or prolix expression. Specifically: a. Assume the richest temporal, causal and referential connections between described situations or events, consistent with what is taken for granted. b.

Assume that stereotypical relations obtain between referents or events, unless this is inconsistent with (a).

c.

Avoid interpretations that multiply entities referred to (assume referential parsimony); specifically, prefer coreferential readings of reduced NPs (pronouns and zeros).

d.

Assume the existence or actuality of what a sentence is about (if that is consistent with what is taken for granted).

This principle is said to cover a whole range of implicatures, e.g., conditional perfection, conjunction buttressing (asymmetric coordination), bridging, inference to stereotype, negative strenghtening, preferred local coreference, mirror maxim, nominal compounds, specialization of spatial terms, and possessive interpretations. Let us discuss the case of conjunction buttressing, a case that has been analyzed as an observation of the fourth sub-maxim of Manner Be orderly! in Grice’s framework (Posner 1980; Carston 2000). Consider the utterance p = Anna turned the switch and the motor started, giving rise to the implicature q = ‘Anna turned the switch and then/therefore the motor started.’ Again, we distinguish the following cases:

132   

   Defining lying

(a) p is true and q is true: It is true that Anna turned the switch and the motor started, and that the starting of the motor followed or was caused by the turning of the switch. Here, no lie is involved. (b) p is true and q is false: The motor did not start because of Anna’s turning the switch, or the motor started simultaneously with Anna’s turning the switch, and the speaker knows this. In this case we have a lie. (c) p is false and q is true: If it is false that Anna turned the switch and the motor started (it is sufficient if one of the two conjuncts of p is false), then it cannot be true that the motor started because of Anna’s turning the switch, or after she did so. This is the case because temporal or causal ordering requires that the relevant events exist in the first place. Hence, the case of p being false and q being true cannot occur. (d) p is false and q is false: It is false that Anna turned the switch and the motor started. So we have a lie. Moreover, the implicature that Anna’s turning the switch is the cause of the starting of the motor or followed by it, is false, too, because the relevant events do not exist. So this case is a lie. A further case of interest may be characterized as ‘insinuation’ (Vincent and Castelfranchi 1981). If A boasts My father works for the FBI and B replies And my father works for the United Nations, B wants to suggest that her/his father is a diplomat, though in reality, he works for the cleaning service. This is in line with the I-principle: If it is possible to build on stereotypical assumptions, the speaker should do so (Levinson 2000: 112–122). Now we turn to the M-Principle (Levinson 2000: 136–137): (33)

M-Principle Speaker’s maxim: Indicate an abnormal, nonstereotypical situation by using marked expressions that contrast with those you would use to describe the corresponding normal, stereotypical situation. Recipient’s corollary: What is said in an abnormal way indicates an abnormal situation, or marked messages indicate marked situations, specifically: Where S has said “p” containing marked expression M, and there is an unmarked alternate expression U with the same denotation D which the speaker might have employed in the same sentence-frame instead, then where U would have I-implicated the stereotypical or more specific subset d of D, the marked expression M will implicate the complement of the denotation d, namely d̅ of D.

Lying and falsely implicating   

   133

Note that only the first (Avoid obscurity of expression) and the third (Be brief (avoid unnecessary prolixity)) sub-maxim of the Gricean maxims of Manner survive in Levinson’s M-Principle. Levinson views the second sub-maxim Avoid ambiguity in connection with generality narrowing, and this phenomenon is subsumed unter the Q-principle (Levinson 2000: 135). The fourth sub-maxim, Be orderly, is not needed anymore because the notorious cases of conjunction buttressing fall under the I-principle in Levinson’s framework. Moreover, Levinson (2000: 135) notes the general cognitive status of this general semiotic principle of linearization, and he questions its status as a maxim. The M-principle is supposed to cover a range of cases, among them lexical doublets and rival word formations, nominal compounds, litotes, certain genitive and zero morph constructions, periphrasis, repetition, and reduplication. As far as I can see, many of these cases may be explained in terms of the Q- or I-principle; in other cases, it is not clear at all that we have “the same denotation” required in the Recipient’s corollary of the M-principle, thus questioning whether a separate M-principle is needed at all (Meibauer 1997). With these reservations in mind, let us take the case of periphrasis and apply our diagnostics. Let us assume that p = Bill caused the car to stop, and q = ‘He did this indirectly, not in the normal way, e.g., by use of the emergency brake.’ (a) p is true and q is true: No lie is involved. (b) p is true and q is false: It is the case that Bill caused the car to stop, but he did it in the normal way. Therefore, we have a lie. (c) p is false and q is true: Bill did not stop the car at all, and the speaker knows this. Then it cannot be the case that he did not stop the car in a normal way. Thus, this case must be excluded. (d) p is false and q is false: Bill did not stop the car, and the speaker knows this. Moreover, it is not the case that he stopped the car in an abnormal way simply because he did not stop the car at all. Therefore, we have a lie. In (34), our findings are summarized. In the first table, you find a survey of types of GCIs and cases. In the second table, the results for case (c) “p is false and q is true” are shown. (34)

Lying in standard cases of GCI

Q-impl. (scalar) Q-impl. (clausal) I-impl. M-impl.

(a) p true, q true

(b) p true, q false (c) p false, q true (d) p false, q false

No lie No lie No lie No lie

Lie Lie Lie Lie

Not possible Not possible Not possible Not possible

Lie Lie Lie Lie

134   

   Defining lying

Case (c): p is false and q is true [SIT = situation, ASS = assertion p, IMPL = GCI q] SIT All colleagues were not drunk (=sober). [scalar] ASS Some colleagues were drunk. If all colleagues were not drunk (=sober), then it cannot be true that some colleagues were drunk. Thus p is false.

IMPL ‘Not all colleagues were drunk.’ If all colleagues were not drunk (=sober), then it cannot be true that not all colleagues were drunk (=sober). Thus q cannot be true.

SIT The doctor does not believe that the patient will not recover. [clausal] ASS The doctor believes that the patient will not recover. If the doctor does not believe that the patient will not recover, then it cannot be true that the doctor believes that the patient will not recover. Thus p is false.

IMPL ‘The doctor may or may not know that the patient will not recover.’ If the doctor does not believe that the patient will not recover, then it cannot be true that the doctor may or may not know that the patient will not recover. (The doctor is not in the state of epistemic uncertainty.) Thus q cannot be true.

SIT It is false that Anna turned the switch and the motor started. ASS Anna turned the switch and the motor started. If it is false that Anna turned the switch and the motor started, then it cannot be true that Anna turned the switch and the motor started. Thus p is false.

IMPL ‘Anna turned the switch and then/ therefore the motor started.’ If it is false that Anna turned the switch and the motor started, then it cannot be true that Anna turned the switch and then/therefore the motor started. Thus q cannot be true.

SIT Bill did not stop the car. ASS Bill caused the car to stop. If Bill did not stop the car, then it cannot be true that Bill caused the car to stop. Thus p is false.

IMPL ‘He did this indirectly, not in the normal way, e.g., by use of the emergency brake.’ If Bill did not stop the car, then it cannot be true that he did this indirectly, e.g., by use of the emergency brake.’ Thus q cannot be true.

In sum, then, it seems that the third condition, namely “p false, q true”, is ruled out, at least for the cases considered here. Thus we have found a fundamental difference to the case of PCI exemplified by the Story of the Mate and the Captain where we found that it may very well be the case that p = Today, October 14th, the Captain is not drunk is false, while q = ‘The captain is usually drunk’ is true. Obviously, the difference has to do with the fact that in GCIs, we have an inherent relationship between p and q mediated through pairs of lexemes that figure in the derivation of the implicature, whereas in the case of PCI the relationship between p and q is contextually induced. Whether false GCIs or false PCIs are “closer” to a lie is not easy to judge. On the one hand, one may argue that false GCIs are “closer” to a prototypical lie than

Lying and falsely implicating   

   135

false PCIs (as Sweetser 1987: 60 did), an impression that I share with regard to cases like the following. When a mother asks her child How many cookies did you take? and the child answers One, thereby implicating ‘Not more than one’, this will count as a lie, if untrue. On the other hand, it may be argued that PCIs are harder to cancel and more difficult to calculate than GCIs, so that lying with GCIs is not as salient as lying with PCIs. At least in exchanges like in (21), it seems that false PCIs are closer to a lie than false GCIs. As far as I can see, this has to do with the greater relevance of the PCI than the GCI with regard to A’s question. After all, A wants to know what the time is and where John is, and not, whether some of the guests are already leaving. If this line of reasoning is on the right track, it follows that the just mentioned case in which the scalar implicature is highly relevant for the questioner is to be judged differently. Many GCIs seem to be derived automatically whereas the derivation of PCIs demands more effort of calculation. What matters is the amount of false information in relation to the questioner’s “purpose or direction of the talk exchange” in which they are engaged. The lying potential of an information rises in relation to its relevance in the utterance situation. Let us take stock: I have argued that implicatures, as additional propositions, are bound to the utterance of the speaker. They are thus part of the total signification of the speaker’s utterance (cf. Grice 1989b: 41). Mismatches and misunderstandings aside, implicatures are normative. It is thus plausible that, even if the notion of implicature demands cancellability in a possible context, some implicatures are quite hard to cancel in the situation of an utterance. This seems to be the case especially with PCIs. If an implicature is not believed by the speaker but nevertheless presented as true, the speaker lies by conveying this implicature. The classical notion of lying that is restricted to untruthful assertion is thus considerably enlarged in scope. Moreover, we have seen that lying with GCIs is also possible. But with GCIs, it is not possible to lie when p is false and q is true. This result underpins the significance of the GCI-PCI-distinction. While acknowledging the case of untruthful implicature, Dynel (2011) insists that intentionally false (= untruthful) implicating should be considered as deception, not lying. I have no quarrel with this opinion. The point is that we highlight the intimate connection between assertion and lying if we include intentionally false implicatures in our definition. Why this is wrong Dynel (2011: 159–162) does not really explain, as far as I see. All that she presents is an argument from authority (“ipsedixitism”): “Whilst the notion of untruthful implicature appears to be correct, that of an implied [implicated, JM] lie is not convincing, because a lie, as traditionally defined by most authors, is a type of deception and an untruthful proposition, here what is said, not untruthful implied [implicated, JM] meaning (save lies based on tropes used in assertions […]).” This move misses

136   

   Defining lying

the point because untruthful implicatures are always bound to assertions, as I have pointed out at length. In addition, since in a Gricean framework tropes are interpreted on the basis of maxims, this seems to prove the point that lying while using false implicatures is a case of lying. As said before, Saul (2012) introduces the notion of misleading as opposed to lying. However, a clear definition of misleading is lacking (see also Stokke 2013c). According to Saul (2012: 71), there are two main characteristics of misleading versus lying. First, “‘misleading’, unlike ‘lying’, is a success term”. This means that S would only mislead H if H was successfully duped. On the contrary, if H does not believe the liar, s/he is nevertheless belied to. Second, misleading may be accidental, whereas lying must be deliberate. Saul illustrates misleading by telling examples. One standard example is Clinton’s famous utterance There is no improper relationship, made in the course of the Lewinsky trial. This utterance was literally true at the time of speaking. However, it was intended to implicate either that there has never been a sexual relationship, or that there was no vaginal intercourse. Be this as it may, these inferences are easily explained by the operation of the I-principle, in Levinson’s terms (see 32). Another example given by Saul (2012: 33) is (35a): (35)

a.

Billy went to the top of the Empire State Building and jumped.

b.

Billy went to the top of the Empire State Building and [then] jumped [off the edge of the building].

The scenario is that Billy jumped up and down three times. So the enrichments in (35b) are false. Saul (2012: 34) comments: “Intuitively, then, my utterance was true but misleading: I neither lied nor accidentally said something false.” However, the notion of misleading thus boils down to the concept of being not explicit enough to avoid misunderstanding. What is more interesting, I think, is the case in which someone uses (35a) to deliberately lead the hearer into thinking that Billy wanted to commit suicide. And this is our case of saying something that is literally true while implicating something that is false. If the cases discussed by Saul can be analyzed with the help of the deliberately false implicature, one should do so because the concept of conversational implicature is a well-established pragmatic concept. Misleading is then restricted to something that is a by-product of misunderstanding (see Schlesinger and Hurvitz 2008; Meibauer 2013d). In this section, I have argued at length that an extension of the notion of lying in order to capture the aforementioned boundedness of the intentionally false implicature to an assertion that could be true makes sense. In the next section,

Falsely presupposing and falsely conventionally implicating   

   137

we will go into the related question of an intentionally false presupposition or conventional implicature.

4.6 Falsely presupposing and falsely conventionally implicating It is tempting to further extend the concept of lying so that it includes falsely presupposing. This is no simple task, since the concept of presupposition, as well as its delimitation against the notion of conventional implicature, has always been a matter of dispute. Hence my aim is of a more modest nature: to point out what one would gain from such an extension and what the risks are. Recall that Vincent and Castelfranchi (1981) already had the type ‘precondition or presupposition faking’ in their taxonomy of deception. Let us take up our example from section 3.8 in which the S-family test was applied. (36)

a.

Ken has quit dealing.

b.

Ken has not quit dealing.

c.

Has Ken quit dealing?

d.

If Ken has quit dealing, Barbie would be glad.

Triggered by the verb to quit s.th., all of these sentences presuppose that Ken was dealing. Chierchia and McConnell-Ginet (1984) assume that this information is part of the background knowledge of interlocutors. Presupposition-triggering elements thus serve to make these background assumptions clear in the development of the ongoing discourse. Now if the speaker of these utterances knows for sure that it was not the case that Ken was dealing, s/he may be accused of lying. To intentionally convey a false presupposition appears to be as much lying as intentionally false implicating (if one accepts the latter concept). Potts (2005) introduces the notion of conventional implicature as a proposition that survives the S-family-test but is distinct to presuppositions because it is not backgrounded, cancellable, and cannot be ‘plugged’ (as for the cancellability of presuppositions, consider classical examples like I haven’t stopped smoking. I’ve never smoked). Conventional implicatures are, like the so-called at-issue entailments (semantic entailments), asserted while both conventional implicatures and presuppositions usually pass the negation test (Jary 2010: 134–137).

138   

   Defining lying

(37) S-family test is backgrounded is cancellable can plug

Presupposition + + + +

Conventional Implicature + – – –

It follows that according to this classification, conventional implicatures sensu Potts are assertions, hence could be used for lying. For instance, if the speaker of (38)

Ken, a convicted dealer, made a million within four weeks.

knows that Ken is not a convicted dealer but wants their addressee to believe that he is, then the speaker is lying by using the conventional implicature that Ken is a dealer. This is not to say that the concept of conventional implicature is the only one to capture the respective meaning effect. Jary (2010: 134–149) argues at length that the CI-tests could be used as a test for the ‘main-point status’ of a proposition. The supplements discussed by Potts (2005) are then propositions which do not contribute to main-point status. As for presuppositions one might further extend the definition of lie along the following lines: (39)

Lie – extended definition (PSP) S lied at t by uttering the declarative sentence σ iff (a) the definition of the lie holds, (b) or S thereby presupposed that q, but actively believed that not q.

In this version, the presuppositions, as additional propositions becoming relevant in the discourse, are triggered by the declarative sentence. However, if other sentence types are included, it is possible to lie while, e.g., using interrogative or imperative sentences. Thus, the utterance Who saw the red car? presupposes that there is a red car. However, when the police officer knows for sure that there is not a red car involved in the crime, but nevertheless asks this question, they are liable to try to deceive their audience, and, according to (39), are lying. The approach just sketched fits nicely into the recent work of Zhang (2008). Dealing with pragmatic presupposition, Zhang (2008) criticizes Simons’ (2004) approach which ranks relevance (in the sense of Relevance theory) as the most important feature of presuppositions. Instead, Zhang develops the view that truthfulness of the participants is the most important parameter in face-to-face conversation.

Falsely presupposing and falsely conventionally implicating   

   139

Let us consider the husband case discussed in Simons (2004), and taken up by Zhang (2008: 370–375). The scenario is that S says I met Jill’s husband yesterday. In the situation, “the hearer knows that Jill is single and infers that the speaker wrongly takes someone as Jill’s husband”. (40)

I met Jill’s husband yesterday. Step 1: The pronoun I refers to the speaker; the deictic knowledge Kd of S and H matches. Step 2: It is conventionally acceptable that someone has a husband given that she is a woman; the linguistic knowledge Kl of S and H matches. Step 3: The current knowledge Kc of S and H does not match: The speaker assumes “Jill is a woman; Jill has a husband” while the hearer assumes “Jill is a woman; Jill is single”.

In this situation, “when the participants hold contrasting knowledge about the presuppositional proposition, competition may occur between them, and as a result either accommodation or cancellation may follow” (Zhang 2008: 373). In a variant of this case the speaker makes the wrong assumption that Jill has a husband and the hearer, being ignorant of Jill’s marital status, may take this proposition for granted and accommodates it. Zhang (2008: 374) concludes that “it is fair to say that not all actions of accommodation by the hearer are based on a safe ground, and not all agreements achieved in this way are successfully compatible with the truth conditions of the actual world”. Thus, it is not the truth that matters with respect to presuppositions, but the “truthful commitment of the participants” (Zhang 2008: 375). However, the typical assumption of the hearer that the speaker is truthfully committed to the presupposition may be wrong. To deliberately bring a false presupposition into play may be a refined strategy of lying. Like in the case of untruthful implicature, the cognitive burden is the hearer’s because they were confident in the speaker’s honesty and trusted them. Zhang (2008: 376) proposes a principle dyad called “Principle of truthful presupposition” that has a speakerand a hearer corollary: (41)

Principle of truthful presupposition PrincipleS: “Presuppose only propositions that you have knowledge about, for truthfulness’s concern.” PrincipleH: “Accept whatever is presupposed by the speaker, for truthfulness’s concern, except when your knowledge forcefully disproves it.”

140   

   Defining lying

While this principle echoes Grice’s maxim of Quality, when related to presuppositions, the hearer’s belief in the speaker’s truthfulness and cooperation may be unwarranted. While presuppositions, like implicatures, are propositions added to the proposition(s) contained in the utterance, it is clear that they are not bound to special sentence types. Quite on the contrary, the S-family-test shows that they are compatible with different sentence types. Does this pose a problem for our approach to defining lying as insincere assertion, as assertions are connected to declarative sentences? I think it is possible to maintain that the prototypical case of lying is the one in which the liar asserts something which they do not believe. Assertions, in turn, are prototypically realized by declarative sentences. If deliberately false presuppositions are brought into connection with lying, be they triggered by declarative sentences or by other sentence types, we surpass the restriction to assertions and acknowledge the insight that it is also possible to lie while using, e.g., interrogatives like Did you see Ken’s Moto Guzzi? while assuming that he does not own such a bike.

4.7 Underdeterminacy and explicature/impliciture In truth-conditional semantics, the propositional content of a sentence gives rise to truth-conditions. This is the usual view of semantic minimalism. The only access to the context of an utterance is via indexical variables. There are, however, several other phenomena that point to a systematic underdeterminacy of propositions. This is a major concern of the contextualists who developed a number of approaches to deal with the facts. Our perspective is clear: We will ask whether it is possible to lie while using explicatures/implicitures. A very useful overview over of the crucial phenomena is given by Carston (2002: 19–30) who draws on work by Gross (2001); for a summary, see (2) in section  1.7. Standard examples (beyond the resolution of ambiguities, indexicality assignment, and cases of ellipsis like On the top shelve as an answer to a preceding question) include the cases of missing constituents (42), unspecified scope (44), underspecificity/weakness of constituents (46), and overspecifity/ narrowness of constituents (47). Let us start with missing constituents: (42)

a.

Paracetamol is better.

[than what?]

b.

It’s the same.

[as what?]

c.

She’s leaving.

[from where?]

Underdeterminacy and explicature/impliciture   

d.

He is too young.

[for what?]

e.

It is raining.

[where?]

   141

The materials in brackets are ‘missing constituents’, i.e., constituents that must be supplied pragmatically. Only by completion of the utterance is it possible to judge it as true or false. Note that Carston (2002: 22) adds example (38d) that we already know from Davidson’s account in section 3.7. This nicely points out the basic question of how to arrive at completing the underdetermined proposition. The following set of examples provides arguments for the context sensitivity of the predicative adjectives (Carston 2002: 23; examples taken from Gross 2001): (43)

a.

Bob is well groomed.

(scalar)

b.

The fruit is green.

(part-dependent)

c.

That is difficult.

(relative)

d.

It is serviceable.

(vague)

Consider (43a): Bob may be well groomed for an average student of German philology, but not for a job at the Deutsche Bank. Thus, the adjective is scalar in that “it allows for comparison between things with respect to the degree with which they have the property concerned” (Carston 2002: 23). In cases like (43b) it is relevant for the truth or falsity of the utterance which part of the apple is green, for example, the interior, the peel, or the stem. Hence, the meaning of the adjective is ‘part-dependent’. Note that the relevant part may be supplied by the context of the utterance. Similarly, in (43c) it is crucial to whom or what the predicate is applied. What is difficult for Ken might not be difficult for Barbie. Therefore, this type of predicate may be called ‘relative’. Example (43d) shows us a ‘vague’ predicate because the standards of precision for the application of the meaning of serviceable may vary from context to context, as Carston (2002: 24) puts it. To demonstrate that all four kinds of context-sensitivity may play a role in the interpretation of predicates, she gives a more elaborate version of (43d) in (43d'): (43d') This program is serviceable, according to such and such a standard, compared to other programs for such and such a task, for beginners in computing.

142   

   Defining lying

The next set of examples has to do with ‘scope ambiguities’ (and thus may be subsumed under the cases of ambiguity resolution). In each case, the negation element not is involved. (44)

a.

Everyone is not hungry.

b.

She didn’t butter the toast in the bathroom with a knife.

c.

The local witch didn’t put a spell on us.

d.

[Bill ate some cakes.] Bill didn’t eat some cakes; he ate all of them.

The utterance (44a) has two readings: The first reading may be paraphrased as ‘not every one is hungry’ (= some are hungry, others are not), the second reading as ‘no one is hungry’ (it is true for every single person that they are not hungry). Thus, in the first reading, the negation element takes scope over the quantifier; in the second reading the quantifier takes scope over the negation element. Example (44b) demonstrates that the negation particle can take different scope over the single constituents. In spoken language, this can be made clear by focussing single constituents as in (44b'): (44b') She didn’t butter the TOAST in the bathroom with a knife. She didn’t butter the toast in the BATHroom with a knife. She didn’t butter the toast in the bathroom with a KNIFE. Example (44c) has to do with the assumption whether there exists a local witch. Dependent on this assumption, two readings are available: First, there is a local witch, but she did not put a spell on the referents. Second, a spell had not been put on the referents, because a local witch did not exist at all (‘the local witch didn’t put a spell on us; there is no witch around here’). Of course, this latter example works when cancelling the existential presupposition triggered by the definite noun phrase the local witch. Example (44d) is usually discussed under the heading of ‘metalinguistic negation’. If it is true that Ben ate all the cookies, then it is also true that he ate some of them. Despite this true entailment there is no contradiction here, because some cookies is taken over or quoted from the utterance in brackets. The expressions form a Horn scale where some +> not all is the typical implicature. Hence, (44d) is a case in which a scalar implicature is explicitly cancelled.

Underdeterminacy and explicature/impliciture   

   143

Though the examples in the following set are, according to Carston (2002: 26), “fully propositional”, she nevertheless argues for a pragmatic process that “is required to arrive at the proposition intended by he speaker”. (45)

a.

Mending this fault will take time.

b.

The north island is some distance from the south island.

c.

Something has happened.

d.

I haven’t eaten lunch.

e.

I haven’t eaten frogs’ legs.

f.

There’s nothing on telly tonight.

Examples (45a–c) seem to express “a trivial obvious truth”, as Carston (2002: 26) puts it, for all activities take place over a certain time (45a), every island is in a certain distance from another island (45b), and it is always the case that something is happening. What the speaker wants to convey is that ‘mending this fault will take a longer period of time than such fault-mendings standardly take’ (45a), ‘the north island is further from the south island than you think’ (45b), and ‘something bad has happened on the day of utterance [to x]’ (45c). Examples (45d) and (45e) have to be temporally specified; but whereas (45d) is probably intended to convey I haven’t eaten lunch today, (45e) probably amounts to I haven’t eaten frogs’ legs in my life. This shows the context-dependency of the transitive construction with the verb in the present perfect. (45f) is usually understood as meaning that there is nothing of interest for the speaker on television. Particularly in those cases that contain quantifying elements like something (45c) or nothing (45f), it is tempting to argue that the respective sentences contain a hidden variable that has to be contextually filled. If this line of argumentation is on the right track, it would not be necessary to assume special pragmatic processes that do the job. Finally, Carston (2002) discusses two further sets of examples. These examples, she points out, seem to involve “not the adding of conceptual constituents, but rather adjustments to linguistically encoded concepts” (p. 27). The first set shows ‘underspecificity or weakness of encoded conceptual content’: (46)

a.

I’m tired.

b.

Anna wants to meet a bachelor.

c.

The path is uneven.

144   

   Defining lying

In example (46a) it has to be made clear what the relevant degree of tiredness is; in (46b), one would like to know whether any bachelor will do for the meeting, or only one who wants to get married, and in (46c), one would like to know whether the road is uneven for a bicycle or a Land-Rover. The relevant concepts should be narrower or strengthened in order to arrive at the full speaker-meaning. The second group of examples shows the reversed effect, namely overspecificity or narrowness of encoded conceptual content. (47)

a.

Her face is oblong.

b.

The steak is raw.

c.

The room was silent.

The literal meanings of oblong, raw, and silent are too narrow, they must be widened or made looser. The face in (47a) is not exactly oblong, the steak not totally raw, and the room not completely silent. It could be argued, of course, that these examples are cases of metaphors. Metaphors are treated by Carston as ad hoc-concepts. The question, of course, is whether all these wider aspects of interpretation are a proper part of the (extended) semantics of the linguistic elements in question, or whether they are a sort of trigger for pragmatic processes that supply the right kind of determination. Carston (2002), so much is clear, aims at the (Relevance theoretic) idea that pragmatic processes lead to so-called ‘explicatures’ which are developments of logical forms. For example, in the case of missing constituents she holds “that pragmatic processes can supply constituents to what is said solely on pragmatic grounds, without any linguistic pointer” (Carston 2002: 23). There is a competing notion in the literature, namely ‘impliciture’, as coined by Bach. Bach (2010) explains the commonalities and differences between the Relevance theoretic notion of explicature and his own notion of ‘impliciture’. Partly, this notion is coined because the material to be enriched is implicit in the utterance, not explicit, as the notion explicature might suggest. (Note that ‘explicature’ is meant to be a cognate of to explicate, i.e., to make something explicit that is implicit.). For Bach, the impliciture is not a “development of a logical form”, as the Relevance theorists would have it, but either an expansion (if the utterance has already propositional character but is in need of further specification) or a completion (if the utterance has not yet the status of a full proposition). He argues that some of the Levinsonian GCIs are, in fact, implicitures, for instance, Jack and Jill are married [to each other] or Bill insulted the boss and [as a result] got fired, and in this respect he sides with Carston (Bach 2010: 127). Another point of agreement is a sceptical attitude against the Gricean notion of ‘what is said’ because

Underdeterminacy and explicature/impliciture   

   145

this notion not only implies that the semantic content of a sentence must be a proposition, but also forces the assumption that non-literal meaning is a case of ‘making as if to say’, as Grice puts it (Bach 2010: 128). But there are many points of disagreement, all having to do with the Neo-Gricean approach taken by Bach visà-vis the Relevance theoretic approach taken by Carston. Since it is not our main task to compare these approaches here in any detail, it shall suffice to point to Bach’s harsh verdict on “relevance theorists’ highly idiosyncratic and misleading use of the term ‘relevance’” (Bach 2010: 135) and the vagueness of the core principles of Relevance theory which do not seem to be very helpful with the problem of “how to quantify and measure degrees of cognitive effects and degrees of processing effort” (Bach 2010: 136) (cf. Noveck and Sperber 2007). According to Carston, explicatures are cancellable. This assumption is, however, highly problematic since cancellability is a defining property of conversational implicatures and the danger of blurring the distinction between explicature and implicature is lurking here. Capone (2009) argues at length that explicatures are not cancellable. His main point is that, like in the related case of PCIs, the explicature is speaker-intended, and hence every attempt at withdrawing the expressed intent leads to implausibility. A similar view has been defended by Burton-Roberts (2010). I think however, that at least in some cases explicatures are cancellable. For example, in You are not going to die [from this little wound] because you are immortal anyway. Furthermore, what is neglected is the distinction of cancellability by an additional statement and cancellability by a possible context (i.e., context-variability). Thus, it appears that explicatures are context variable. For instance, in a standard example like I haven’t had breakfast, there are possible contexts (e.g., the utterance is made by an alien) in which the literal meaning ‘I never had breakfeast in my whole life’ would be evoked. One major problem with the concept of explicature/impliciture is how to fix the norm of a complete or full proposition. For instance, as Cappelen and Lepore (2004: 64–65) argued, even in relation to a seemingly innocent assertion like John went to the gym, setting the stakes high, it could be asked whether the intended full proposition is, for instance, ‘John went to the vicinity of the gym’ or ‘John went into (inside) the gym’. If it is possible to shift the standards of truth-evaluability ad libitum, how can we be sure that a full proposition is ever accessed (cf. Capone 2009: 65–67)? The usual answer is that context knowledge decides how far we must go and that the kind of adequate enrichment is steered by considerations of what a full proposition should look like. In contrast, in the case of conversational implicature it matters what kind of additional information is conveyed. Another problem rarely discussed in the contextualist literature is to what extent the putative examples of explicatures may be considered as cases of ellip-

146   

   Defining lying

sis. The properties listed with respect to (42) match with properties typically ascribed to ellipsis. For instance, the case of missing constituents might turn out as an ellipsis if a certain part of the complete sentence has to be reconstructed from the context, e.g., Paracetamol is better [than Aspirin]. Capone (2009: 74–75) accepts the following list of properties provided by Grundy (2000: 53) and adds further characteristics: (48)

(49)

Explicature (Grundy 2000: 53) a. Explicatures amount to constitutive aspects of what is explicitly said. b.

Explicatures are not linguistically encoded but have to be pragmatically expressed.

c.

Speakers are committed to the explicature of an utterance.

d.

Explicatures are part of what is communicated and, thus, are overtly endorsed by a speaker.

e.

Explicatures are motivated by the indeterminacy of language.

“[…] explicatures are inferences that partially use linguistic meaning and partially use contextualization clues in order to determine the speaker’s unequivocal intention. An explicature is, therefore, the reconstruction of an intention on the basis of what a speaker says in communication in response to the need to reach a full proposition. The full proposition reconstructed by the hearer is one that cannot be cancelled explicitly and one that can be distinguished from implicatures due to the fact that implicatures are additional increments on top of explicatures. Explicatures are more fundamental increments, as without them the proposition expressed would be either contradictory or false or communicatively inefficient even at the sentential level. Explicatures are explicit in that they cannot be denied.” (Capone 2009: 74–75)

If we accept this picture and stick to the standard examples provided by Carston, Bach, and others, we can ask whether it is possible to lie while using explicatures. Suppose the following scenario. Two men, Paul and Pete, are on a dangerous expedition through the jungle. As it happens, one of them, Paul, is bitten by a venomous snake. Pete – but not Paul – knows that this bite will inevitably lead to the death of Paul. Now Pete says You are not going to die [from this little wound]. The material in brackets – the explicature – is intended by Pete and correctly enriched by Paul, and Paul does not entertain the thought that Pete wants

Underdeterminacy and explicature/impliciture   

   147

to tell him that he is immortal. But what Pete has done (with honest intentions, of course) is lying since he knows for sure that Paul will die from his little wound. Similarly, suppose Ken is meeting Barbie who invites him to have lunch with her. When asked, he answers I haven’t eaten lunch [today] (= 45d). The explicature is correctly derived by Barbie, but as a matter of fact, Ken lied to her since at the same day he already had lunch with Lissa. The alternative that Ken wants to assert that he never has eaten lunch throughout his life is excluded because this is not plausible. Note that lying while using I haven’t eaten frogs’ legs [in my life] (45e) does not work, since [in my life] entails [today], and thus the utterance without the explicature would suffice for lying. Elaborating on the notion of ‘what is said’, as used in contemporary debate, Saul (2012) proposes a notion of lying that includes cases of propositional enrichment: (50)

Lying (Complete) (Saul 2012: 65) If the speaker is not the victim of linguistic error/malapropism or using metaphor, hyperbole, or irony, then they lie iff (A) and (B) holds: (A) (1) They say that P; (2) They believe P to be false; (3) They take themself to be in a warranting context. (B) (1)

They say something indeterminate across a range of acceptable complete propositions, CP1…CPn); (2) for each complete proposition in the range CP1…CPn, they believe that proposition to be false; (3) They take themselves to be in a warranting context.

This definition captures the idea that deliberately false, complete (i.e., enriched) propositions are part of lying. This is in line with the above sketch. However, if enrichments of what is said (alias explicatures or implicitures) are part of the concept of lying, one wonders why deliberately conversational implicatures and presuppositions should not be part of the overall concept of lying. We may summarize, then, that explicatures/implicitures, understood as expansions or completions, or more generally, as enrichments of underdetermined propositions, appear to also be usable in the act of lying. This is, however, speculative at the moment since the empirical research into explicatures/implicitures is in its earliest stages, and the very notion of explicature/impliciture as developed in (48) and (49) is not yet crystal-clear.

148   

   Defining lying

4.8 Vagueness and imprecision In (43d), we already had a case of a vague adjective. Vagenuess, in general, is a meaning phenomenon where pragmatic enrichment, with the aim of making a proposition more complete or more precise, is difficult if not impossible since it is the vague meaning that is intended. Nevertheless, even vague meanings are evaluated with respect to truth, and there are many approaches that try to fix vague meaning by admitting degrees of truth. However, as van Deemter (2010: 284) states: “If degrees of truth are adopted, the question whether a statement is truthful is no longer clear-cut. If a real-estate advert describes the puny garden of your house as ‘sizeable’, for example, is this a lie or a mere exaggeration?” Vagueness, then, is another meaning phenomenon that invites manipulation of the truth and lying. Channel (1994), in her descriptive study of English vague terms, already hinted at using vague expressions for “deliberately withholding information” and “using language persuasively”. For the former case, see (51): (51)

a.

She’s about 29 years old.

b.

… I’m not talking about the actor who plays him, Michael whatsisname. He’s hardly in the movie anyway …

Utterance (51a) is made in a context in which the speaker knows that the referent is 30. Presupposing that it is more flattering to have an age “under thirty”, the addressee is deceived (the speaker is not as informative as they could be), while the speaker is, strictly speaking, not untruthful because the approximator about appears to be compatible with one year more or less. In (51b), the speaker knows what the surname of the actor is, but they prefer to produce the dummy expression whatsisname to convey a pejorative attitude. Again, the speaker is underinformative and thus deceives the hearer relative to the norm of providing complete information. As for “using language persuasively”, Channel (1994: 179–180) mentions cases in which numbers are rounded in order to make statistics more appealing or to create the impression of bigger success. To be imprecise where precision is called for (and to be precise where imprecision is adequate) is a linguistic possibility that invites manipulating the truth. In the following, I will consider three further cases in which it can be argued that imprecision may pave the way for lying. These cases are in turn: (a) truthfulness when telling the time, (b) truthfulness when telling distances, and (c) truthfulness with respect to taste predicates.

Vagueness and imprecision   

   149

The first case relates to truthfulness when telling the time. Look at the exchange in (52): (52)

Peter: Mary arrives at three o’ clock. John: [Mary arriving 15 seconds later]: #You said she came at three!

When S tells H that Mary arrives at three o’clock, but later he finds that she in fact arrives 15  seconds later, it is unreasonable of H to complain (# indicating pragmatic inadequateness). Lasersohn (1999: 522), to whom we owe the example, comments: “But whether or not John is acting unreasonably in this situation, I think we have to concede that he is, strictly speaking, right: when I told him that Mary arrived at three, I said something that was literally false, not true. As people often do, I was just speaking a little loosely. My defense is not that I was telling the truth, but that what I said was ‘close enough’ to the truth for practical purposes.”

With regard to the sentences in (53), Lasersohn (1999: 522) points out a semantic puzzle. If (53a) is literally true, then it seems truth-conditionally equivalent to (53b): (53)

a.

Mary arrived at three o’ clock.

b.

Mary arrived at exactly three o’ clock.

But these sentences do not mean the same, since exactly contributes to the sentence meaning. Words like exactly are viewed as lexical devices signalling the intended degree of the approximation of truth. Sentences like (53a) allow greater ‘pragmatic slack’ in comparison to sentences like (53b). Therefore, lexical elements like exactly are called ‘slack regulators’. Lasersohn (1999: 527) develops a formal theory of ‘pragmatic halos’: “Given an expression α denoting some object x, I like to think of the set the context associates with x as arrayed around x in a sort of circular cluster, so I will call this set, together with its ordering relation, the pragmatic halo of x, extending the terminology, as the pragmatic halo of α.” The second case relates to truthfulness when telling distances. The following examples are taken from Krifka (2002: 430–431). Note that the distance between Amsterdam and Vienna is in fact 965 kilometers. (54)

S: H:

The distance between Amsterdam and Vienna is one thousand kilometers. #No, you’re wrong, it’s nine hundred sixty-five kilometers.

150   

   Defining lying

According to Krifka, in this context H’s answer is true but pedantic, while in (55), it is perfectly adequate. (55)

S: H:

The distance between Amsterdam and Vienna is nine hundred seventy-two kilometers. No, you’re wrong, it’s nine hundred sixty-five kilometers.

This shows that speakers face the problem of ‘precision level choice’: “When expressing a measurement of an entity, choose a level of precision that is adequate for the purpose at hand.” (Krifka 2002: 433) Krifka eventually finds that short expressions and vague interpretations are pragmatically favored. Furthermore, short expressions correlate with vague interpretations, while long expressions correlate with precise interpretations. Recently, the idea of a convention or maxim of truthfulness has come under attack by Relevance theorists. Wilson and Sperber (2002: 583) argue “that language use is not governed by any convention or maxim of truthfulness in what is said.” Instead, relevance is what governs language use, and therefore “expectations of truthfulness – to the extent that they exist – are a by-product of expectations of relevance” (Wilson and Sperber 2002: 584). According to Wilson and Sperber, prima facie evidence against the claim that speakers try to tell the truth are lies, jokes, fictions, metaphors, ironies, and, more generally, so-called loose uses of language such as approximations and sense extensions. Note that van der Henst, Carles and Sperber (2002) make an attempt at verifying this claim experimentally with respect to the question of truthfulness when telling the time. What they found is that subjects strived for relevance and not for truthfulness, and that they tended to rounding up. Here I would like to object. That relevance or the choice of the right precision level plays a role in the above cases does not mean that a maxim of Quality (or some device with the same function) is superfluous. Take irony as an example. If Ken says to Barbie Dan is a fine friend in order to ironically convey the proposition that Dan is not a fine friend then this is nevertheless compatible with his statement being a lie. Ken really believes that Dan is a fine friend, but for some reason he wants to deceive Barbie about his attitude to Dan. I do not see how to explain the deceptive manipulation of contents without any appeal to the notions of truth and truthfulness. In short, lying is not evidence against truthfulness, but for truthfulness. Coming back to the example by Lasersohn (1999) in (53a), I would like to argue that the pragmatic slack opens up another promising avenue for the liar to go, similar to the case of lying while saying the truth working with intentionally false implicatures. Imagine a situation in which it is clear that precision is called

Vagueness and imprecision   

   151

for. When a defendant is asked when Mary arrived they are nevertheless licensed to answer at three o’ clock, even if they know (and know that this matters) that her arriving happened at 3 minutes past three o’clock. The pragmatic slack therefore is therefore also a lying slack. Does this mean that truthfulness is irrelevant? On the contrary, it shows that truthfulness is relevant, and, what is more interesting, that the liar finds niches built into language structure that could be used as ‘lying halos’. As a third case, look at taste predicates like delicious, as discussed extensively in Smith (2010). Suppose, Ken said to Barbie, on the occasion of a romantic dinner, This wine is delicious. Delicious is a taste predicate, so there may be agreement or disagreement between Ken and Barbie. Agreement is most certainly preferred in this situation, and both participants know this and know that the other knows it. As for the disagreement case, the question pops up who is right. Can there be such a thing as “faultless disagreement”, or must one of the participants be the victim of a false judgment? Smith (2010) argues for a relativist approach. Even if both parties agree, he shows, it may be unclear, in what perspective they agree. For instance, (56a) may be understood as (56b-e): (56)

a.

This is an excellent Bordeaux.

b.

This is excellent for a wine and it is a Bordeaux.

c.

This is excellent within the class of Bordeaux wines.

d.

This is excellent for a Bordeaux wine.

e.

This is excellent given how Bordeaux wines ought to be.

Hence, Smith (2010: 213) concludes that “there will be more than one way for [56a] to be made true and speakers will often selectively attend to, or focus on, just one way of making that sentence true and imagine or hear the sentence, when uttered by others, as saying what they take it to mean: i.e., [56b] or [56e].” Along the same lines, when parties disagree, they usually disagree over the right standards with respect to a comparative class. Smith (2010: 214) concludes: “When we switch standards or norms, or weigh them differently, we may get different verdicts on the truth-value of the same statement – a statement made with the help of a gradable adjective. The same statement, making the same claim, would be true relative to one criterion of assessment but false relative to another.” Now suppose that Ken dislikes the wine. He could go on and say truthfully I do not find the wine delicious. This could be the start of an engaged discussion of wines, etc. But this is too risky for Ken, so he chooses to utter This wine is delicious. Then he does lie according to his own norms, but since these norms are

152   

   Defining lying

covert to Barbie, she believes him and there is no disagreement between them (at least not with respect to the wine). As a matter of fact, judgements of taste seem to play an important role in romantic conversations, so these judgments have to be carefully done and developed within the discourse (cf. Barker 2013). The idea, then, is that vague predicates allow for lying without the risk of disagreement.

4.9 Prosocial lies In some analyses, so-called “white lies” are treated as a separate type of lying. Hence, we can ask whether we need to modify the proposed definition in (27), p. 125. In this section, I will argue that white lies are lies according to my definition. White lies are submitted to a normative dimension which is not the subject of a linguistic study. However, they appear to be linked to certain standard contexts. Since speaker’s and hearer’s knowledge about standard contexts are pragmatic in nature, there is a connection between the use of white lies and the use of lies which deserves further exploration. Let us begin with a terminological note. Although the notion of “white lie” is the more popular one, I prefer, together with parts of the literature, to speak of “prosocial lies” in what follows. This term fits better to the basic assumption that there is a difference between antisocial and prosocial lying. The basic cognitive ability of lying is neutral with respect to these dimensions. A look at standard examples of prosocial lying provides us with at least the following types (Dietz 2003; see also Bok 1999): (a) benevolent lying; (b) lying in self-defence; (c) lying in order to protect the private sphere. (a) benevolent lying. As for benevolent lying, several classic scenarios should be mentioned. When a priest is asked by a dying mother Is my son okay?, the priest will lie to her He is well and alive although he knows for sure that the son had a severe car accident yesterday. Similarly, we can think of a doctor who refrains from telling his moribund patient that he will die within 24 hours. In these cases, the liar is benevolent to someone. In both cases, the liar thinks that it is better to tell a false fact than to tell the truth. Telling the truth would massively enhance the mother’s or the patient’s suffering. Whether the lies of parents about Santa Claus are benevolent in the described sense is not easy to decide. In this case, norms are in conflict. On the one hand, parents are benevolent in that they introduce children to certain religious and cultural traditions; on the other hand, this is in conflict with the idea that telling the truth (especially to close relatives) is a high-ranked norm.

Prosocial lies   

   153

The so-called “noble” lie in which the liar asserts that s/he has done something bad what in fact has been done by another person is also a benevolent lie. In this case the liar runs the risk of being punished in order to save the other from punishment. (b) lying in self-defence. When the drug dealer asks me whether I told to the police that he gave crack to the kids, I lie that I never talked to the police. When a murderer comes to my door and asks where my friend Ken is, I will lie to him and tell him that Ken is on holiday, although I know that he is hiding under my bed. In these cases, I lie in self-defence, be it that I want to protect myself against the dealer’s revenge, or that I do not want to be the traitor of my friend Ken. Lying to the drug dealer or to a potential murderer is licensed by the overriding norm that my life and the life of others should be protected. (c) lying in order to protect the private sphere. There are also scenarios in which someone lies because they feel that someone intrudes in her/his private sphere. For instance, Barbie is sadly looking out of the window. When Ken asks What are you thinking?, she answers Nothing special. This a lie because she was just thinking how happy she had been with her former boyfriend Ben. But this is a thought that should not be made explicit, because Barbie wants to protect her private sphere. In addition, it does not seem to be in the interest of both partners to start a fight about jealousy, etc. Hence, Barbie may be acting benevolently, too. These three types of lying scenarios demonstrate that there are prosocial lies: lies that protect social life instead of harming it. I assume that all these cases are covered by my definition of lying. It is important to recall that this definition avoids any mention of moral evaluation. When moral evaluation is taken into account, the manifold reasons for lying are analyzed. These reasons can be bad or good, or can be a complex mixture of good and bad reasons. To present such complex situations, including good liars and bad truth-tellers, is the subject of many famous narratives. Prosocial lies are also possible with untruthful implicatures. Saul (2012) argues that lying is not morally worse than “merely” misleading. (Recall that untruthful implicatures are treated as cases of misleading by Saul.) I agree with this conclusion since it supports my view that untruthful implicatures are a part of the overall signification of an utterance. Yet, the idea that lying is worse than mere misleading has to do with the insight that saying something goes together with more commitment than implicating something. This is also wrong, as I argue, since there must be also a commitment to an implicature (Meibauer 2013d). However, implicatures may be easier to cancel, deny, or clarify. It is easier

154   

   Defining lying

to pretend that the hearer was the victim of a misunderstanding when I deny to have been conveying an implicature in the first place. In principle, it seems possible to include ethical parameters in definitions of lying. Concentrating on a linguistic definition of lying, I refrained from doing so. Yet there is one aspect of moral evaluation that is relevant for pragmatic theory. This is the speaker’s and hearer’s knowledge about standard situations in which it is possibly or even likely that a lie will be presented. Take the following scenario as an example. Barbie has just bought a dress that Ken thinks is awful. She asks him how he likes the dress. Ken answers Great! This is a standard situation of everyday life. Barbie knows that Ken may not speak the truth, yet she refrains from asking him how he really likes the dress. I would like to argue that many situations are well-known to adult speakers as pragmatic situations in which, for several reasons, a lie may be likely. These situations include scenarios in which the liar acts in a prosocial way. Recent experimental results show that speech participants are competent judges with respect to such situations. Levine and Schweitzer (2013) experimentally show that trust in a liar may even increase when a prosocial lie is accepted. The fact that participants are sensitive to the specific kind of evaluation may be seen as evidence for knowledge about typical behaviors relative to such standard situations.

4.10 Conclusions In this chapter, I have argued that a proper definition of lying is possible without a direct reference to the sincerity condition if the maxim of Quality is upheld. The deceptive intent connected with lying follows from intentionally asserting p while believing non-p, and does not need to be expressed separately. Baldfaced lies and cases of self-deception can be accommodated to these definitions. Drawing on the case of intentionally false (untruthful) implicature, it has been argued that  – if the status of implicatures as additional propositions triggered by an utterance is taken seriously – this phenomenon should be captured by an extended version of the standard definition of lying. Thus, cases of “misleading” are not separated from “proper” lying, they are part of the overall act of lying. Walking this avenue further, the cases of falsely presupposing and falsely conventionally implicating were discussed. Notorious problems of defining these notions set aside, we have found evidence for integrating these phenomena into a comprehensive definition of lying, too. Finally, we have discussed the possibility of lying while bringing underdeterminacy phenomena into play. Explicitures or implicitures, as well as vagueness and imprecision, are phenomena in which

Conclusions   

   155

a proposition is in need of enrichment. However, since the context and speaker norms play an important role, these phenomena lend themselves for being used in lying and manipulating the truth. In sum, I do not think that it is always simply truth and falsity that matters in lying, but also the more intricate aspects of modulating the proposition. These aspects are so sublime that in the normal course of a conversation they are hard to grasp; nevertheless, they are there and may develop certain effects in the hearer.

5 Interface Problems 5.1 Introduction In this chapter, we will discuss a number of phenomena that are clearly relevant for the task of locating lying at the semantics-pragmatics interface. All of these phenomena are related to matters of truth and assertion in one way or another, and so the question arises what the respective relation to lying is. For instance, while ironies are not lies, it is nevertheless possible to lie while using an ironical utterance. We will see that the lying potential of assertions is bigger than is usually assumed when subtle effects of implicated propositions are taken into account.

5.2 Aspect-sensitivity, or: Did Clinton lie? There is a nice little controversy about the question whether Clinton did lie, drawing on the question of aspect-sensitivity. While Moore (2000) argues that Clinton did not lie, Saul (2000) argues that Clinton did not lie, but said something false. The subject of this controversy is President Clinton’s utterance at a press conference in January 1998: (1)

I did not have sexual relations with that woman, Miss Lewinsky.

Suppose that “sexual relations” is understood in the broad sense, including all sorts of sexual activities. Then, it seems, Clinton lied. Moore (2010: 250) brings a further supposition into play: “Suppose, further, that Clinton has two names for Monica Lewinsky, and that he is careful to use them in different ways. When he speaks about Monica Lewinsky in her capacity as a White house intern, and about his relations with her while she is acting in this capacity, he calls her ‘Miss Lewinsky’. But when he speaks about Monica Lewinsky as a participant in their improper affairs, he calls her simply ‘Monica’. Suppose also that Clinton was careful not to stand in any improper relation to Monica Lewinsky while she was carrying out her internly duties.”

The point is that the different ways of referring make a difference with respect to the question whether Clinton lied. This has to do with aspect-sensitivity, as developed by Forbes (1997, 1999) and Moore (1999):

Aspect-sensitivity, or: Did Clinton lie?   

(2)

   157

Aspect-sensitivity (Moore 2000: 250) […] if proper names are systematically pegged in this way to different capacities, personae, or, more generally, ‘aspects’ of an individual, these aspects can, in certain cases, figure in the truth-conditions of what is literally said (and not merely communicated) by ‘simple sentences’ (that is, sentences such as (1), that lack any overtly modal, quotational, or psychological ingredients).”

Thus, if the use of Monica Lewinsky refers to other aspects of the respective woman than Monica, Clinton was not lying. This line of argument presupposes a so-called ‘enlightened context’, that is, “a conversational scenario in which all participants know that the sensitive names are, in standard contexts, co-referential” (Moore 2000: 251). Moreover, maxims of perspicuity, “a marked distinction between the way the co-referential names are used, clear links between these distinctive usages and salient aspects of the normal referent, and some conversational reason for wanting to convey information about those aspects” (Moore 2000: 251) are called for in order to convey the intended meaning. When an enlightened context cannot be assumed, it is, according to Moore, nevertheless possible that “the speaker’s utterance says one thing relative to the speaker’s presuppositions (i.e., relative to the ‘speaker’s context’), and something different relative to the audience (i.e., the ‘audience context’)” (Moore 2000: 252). Thus, Clinton did not lie, at least with respect to his own perspective (note the parallel to our discussion of taste predicates in section 7.2), and a sentence like (3) poses no problems: (3)

Monica visits the Oval Office more often than Miss Lewinsky.

In contrast, Saul (2000) holds that Clinton failed to say something truth-evaluable at all. She argues “that simple sentences simply cannot express aspect-sensitive propositions, and that Clinton was wrong to think that he could say something which was aspect-sensitive” (Saul 2000: 256). Moreover, Saul holds that “Clinton’s utterance was simply false if he and Lewinsky did in fact engage in sexual relations, both as he understood ‘sexual relations’ and as that term is understood in English” (Saul 2000: 256). She concludes, then: “If Clinton falsely believed that he was making a true claim about aspects, then he did not lie, because he did not realize that his claim was false. But he nonetheless made a false, aspect-insensitive claim. This solution, I think, does justice to the intuition of falsehood at the same time as it upholds the intuition that such a falsehood would not involve lying.” (Saul 2000: 256)

158   

   Interface Problems

Quite on the contrary, I would like to argue that Clinton was indeed lying. Note the particular phrasing of (1) above where the demonstrative expression that woman is used first, before the name Miss Lewinsky is spoken out. Not only did Clinton speak of “Miss Lewinsky”, he did it in a way that signals distance and a pejorative attitude which becomes clear when it is compared with an alternative utterance like I did not have sexual relations with Miss Lewinsky. I think that this is a deceptive behaviour since it is intended to suggest that Clinton is not acquainted with Miss Lewinsky, let alone having an intimate relationship with her. We may reconstruct this as a conversational implicature (PCI) on the basis of the maxim of Manner. We know that the particular way of phrasing something, as when compared with unmarked alternatives, may give rise to implicatures. And if one accepts my proposal to include untruthful implicatures into the definition of lying, one can very well argue that Clinton lied indeed. The assumption that Clinton might believe that (3) could be true is implausible. Most audiences would conclude from (3) that Monica and Miss Lewinsky are different persons. To prevent this conclusion, one has to be clear about the fact that it is one and the same person, but that visits to the Oval Office where connected with different types of activities.

5.3 Fiction Intuitively, an author who invents a work of fiction that is not true with regard to reality is not untruthful and they are not lying. This is the case because they do not present their fiction as true. It is presupposed that every reader knows the narrated events to be fictional. Fiction, in the prototypical case, may be understood as a text consisting of fictive utterances produced by an author. Not all parts of such a text must be fictive since the author may integrate non-fictive (sincere) utterances. For instance, “Emil and the Detectives” may contain truthful passages related to Berlin as envisaged by Erich Kästner, while all utterances related to the protagonist Emil are fictive. The main problem is, then, how to define fiction. In the following section I will discuss three approaches that are of interest for our leading question concerning the differences and commonalities between fiction and lying. These are the approaches by Searle (1979b), Currie (1990), and Jary (2010); for a detailed discussion of fiction and fictionalism, see Sainsbury (2010). Finally, I will come back to the question of lying authors and characters. Searle (1979b) compares assertions as defined by his system of rules/conditions with their correlates in fiction, since the essential rule, i.e., “the maker of an assertion commits himself to the truth of the expressed proposition” (Searle

Fiction   

   159

1979b: 62), does not hold for the author of a piece of fiction. Instead, the author of fiction “pretends to perform illocutionary acts which he is not in fact performing” (Searle 1979b: 66). However, and this is the crucial distinction to lying, the author does not deceive the hearer; the liar pretends to assert, too, but always intends the hearer to be deceived about their pretence. Refinements are made with respect to third-person vis-à-vis first-person narratives, and with respect to prose fiction versus plays. Take “Emil and the detectives” as an example (translation by May Massee, Doubleday, Garden City/New York): (4)

a.

I might as well tell you – this affair of Emil’s was a great surprise to me. As a matter of fact, I wanted to write an entirely different book: a book in which tigers would gnash their teeth and date palms rattle their coconuts in sheer terror!

b.

In the living room sat a woman with her head bent over the white washbowl. Her hair was undone and hung down like three pounds of wool. Emil’s mother poured the liquid camomile soap over the woman’s blonde hair and rubbed the strange head until it foamed.

In (4a), which is part of the introductory chapter “But the Story Does Not Begin Yet”, the author speaks to his audience. The first utterance in (4a) counts as nonfictive and could be a lie. In contrast, the utterances in (4b) taken from the first chapter “Emil Helps to Shampoo” are, according to Searle, pretences to assert since they are fictive. Note that Searle would assume that both types of utterances are produced by the same author, Erich Kästner. In plays, the author pretends to assert something; but typically he does so because he intends an actor to utter these assertions (and, of course, other illocutions) on stage. It follows that actors do not assert something but also pretend to assert. Hence, actors never lie, at least with respect to their audience. Even when it happens that an actor utters I am hungry and they are hungry indeed, they were not truthful since their assertion was, as part of the play, only pretended. Sainsbury (2010: 9) argues that there is “no need for the idea that a creator of fiction pretends to assert”; instead, he follows Currie (1990) when he suggests “that a creator of fiction is (really) trying to get the audience to engage in makebelieve”. According to Currie (1990), an utterance is fictive if it is made with the intention that a potential audience make-believe that p. P should not be true (or only true by accident). Sainsbury (2010: 8) summarizes:

160   

   Interface Problems

“A fictive intention is one in which the utterer intends a potential audience to make-believe something. When a work results from systematically connected series of utterances some reasonable proportion of which are issued with fictive intentions, and whose truth is at best accidental, it is a work of fiction, otherwise it is not.”

Sainsbury admits, however, that arguments against Searle’s approach are hard to provide. One putative argument is provided by Walton (1990) who argues that certain paintings do not assert, yet are fictions. The other one is proffered by Currie (1990: 17) who mentions a case of someone who mimics “a verbal performance to illustrate an idiotic line of reasoning”; this is not fiction, but the speaker pretends to assert. Both arguments are not convincing; the first, because it is not clear what the propositional content of a picture is, the second, because not every case of pretending necessarily includes fiction. The starting point of Searle’s observations on the “logical status” of fiction certainly is to prove how his definition of assertion is related to fictive assertion. Note that he draws a distinction between vertical conventions, i.e., the usual rules/conditions for an illocution that constitute a connection between the world and the language, and horizontal conventions breaking the vertical conventions, as in the case of fiction. These horizontal conventions are not semantic conventions; they dismiss the speaker from the commitments imposed on them by the vertical conventions. This amounts to construing fiction as a separate language game. However, Searle has not much to say about the specific horizontal conventions being active in fiction. Moreover, the common vertical conventions must still be active since readers must be able to identify an utterance of a fictional character as an assertion or as a lie. If fiction is determined by so-called horizontal conventions, it is only a minor step to assume that “we don’t have to do any pretending to produce or engage with fiction”, as Sainsbury (2010: 11) puts it, simply because it is a separate type of verbal action, similar to joking. “The author need not pretend he is doing anything other than what he is doing, telling a story; and telling a story is simply not the same as pretending to recount facts, for the latter is something liars do and story-tellers may not be liars. The reader need not pretend she is doing anything other than reading the story, entering into it, and, with luck, enjoying it. So pretending seems to lack a special role in a discussion of fiction.” (Sainsbury 2010: 11)

In contrast, make-believe or getting the audience to imagine something is a more plausible property of fiction. Jary (2010: 67) discusses assertion in fiction in connection with cases of ‘discourse embedding’ like the ones in (5):

Irony and tautology   

(5)

   161

a.

Let us suppose the next sentence for the sake of the argument. Snow is white. [supposition]

b.

I therefore put forward the following conjecture: snow is white. [conjecture]

The point is that the sentence Snow is white is not asserted in this context since in (5a) we have a supposition, and in (5b) we have a conjecture. In both cases, there is no “claim that the conditions for asserting that p” obtain. The difference is that a conjecture has to be withdrawn in case the conditions for assertion do not obtain, while this is not the case with supposition. In fiction, neither do the normal conditions for assertion apply nor is there a need to withdraw fictive utterances when not compatible with truth. Jary (2010: 69) sides with Recanati (2000: 49–62) who analyzes fiction as a form of simulation. We have argued that the analysis of fiction as make-believe is more convincing. Fiction is not “simulated assertion” but the author’s undertaking of stimulating the imagination of their audience. What Jary rightly stresses, however, is that assertion, conjecture, supposition, and fiction have the common property of being “an undertaking of commitment (of some kind) to the inferential consequences of asserting that P” (Jary 2010: 70). With respect to fiction, one could argue that there are requirements of consistency, etc., just as if it were normal assertion. Note, however, that there are fictions in which these requirements are stretched, thus inviting different interpretations of the fictive ‘what is said’. The reference to “of some kind” in the above quote is therefore important. When we face an unreliable narrator, the task of the reader is to make sense of the specific representation of a fictive world. On a final note, it should be added that there are numerous cases of lying authors, for instance, authors who lie in autobiographies, and of course, numerous narrations in which lying characters play an important role (see KümmerlingMeibauer and Meibauer 2011 for examples in children’s literature). While Astrid Lindgren (as a third-person author) does not lie, Pippi Longstockings does.

5.4 Irony and tautology In ironical utterances, speakers do not tell the truth. For instance, if I ironically say You are a genius! on the occasion of your spilling the ravioli all over the kitchen floor, then you should understand that I do not literally mean that you are a genius. Quite on the contrary, I want to convey that you are very clumsy. So much is clear: Ironical utterances are not lies since when being ironical the

162   

   Interface Problems

speaker wants the hearer to derive the ironical meaning, while the liar never wants that he hearer detects that they are lying. In this section, we will discuss the Gricean analysis of lying (as well as the Relevance theoretic echo theory and the pretence theory) with the aim of drawing a distinction between lying and irony. Not surprisingly, I will argue that it is possible to lie by using ironic utterances. Furthermore, I will argue that it is even possible to lie with tautological utterances – that is, utterances that are always true. Consider the maxims of Quality. If a speaker S utters X is a fine friend, referring to a person who has betrayed a secret of S’s to a business rival, then the first maxim of Quality is flouted (Grice 1989a: 34). Because it is obvious that S does not believe what s/he says, the hearer reconstructs a related proposition, namely the opposite of p. The ironical implicature qualifies for the status of an implicature because it is calculable, context-dependent, and cancellable. (6)

X is a fine friend. +> ‘X is not a fine friend.’ Calculability: The implicature may be worked out. Context-dependency: There are other contexts in which the implicature is not triggered. Cancellability: X is a fine friend; he was always there when I needed help.

Note that this substitutional analysis – the implicature substitutes the meaning of the utterance – is in contrast with the additive nature of other types of implicature. For instance, the utterance Jones has a beautiful handwriting and his English is grammatical in the context of a reference may have the implicature ‘He isn’t suited very well for this job’ in addition to the propositions that Jones has a beautiful handwriting and that his English is grammatical. Wilson (2006: 1726), in her defence of the echo theory of irony, points out that Grice assumes that in tropes, the speaker is not really saying something; instead, they are merely “making as if to say” something, or they are “purporting to be putting forward” a proposition (Grice 1989a: 34); but if nothing is said and nothing is asserted (which would require a commitment to the truth of the uttered proposition), one wonders why the first maxim of Quality could be violated at all. However, the Gricean approach has been criticized for several reasons: First, the analysis cannot account for ironical questions, requests, and understatements (see the examples in (7)); in ironical questions and requests, the requirement that the speaker should not say what they believe to be false does not play a role, and in ironical understatements, the speaker observes the maxim of Quality while they flout the maxim of Quantity.

Irony and tautology   

(7)

   163

a.

Do you like another glass? [when the hearer obviously is disgusted at the wine]

b.

Listen to this amazing music! [cacophonous noises come out of the radio]

c.

Not quite his best match today! [the player plays awful]

Second, it cannot explain the distinction between irony and metaphor because the latter is also explained with regard to the first maxim of Quality. Thus, uttering You are the cream in my coffee, conveying ‘You are my best friend’, is literally false but it is not ironical. Third, it is not fine-grained enough because it does not follow from He is not a fine friend that he is not a friend at all. Hence, a theory of irony that spells out how the hearer arrives at the exact ironical meaning (e.g., Fritz may still be called a friend, but on several occasions he acted selfishly) is called for. The Gricean approach to irony has been most prominently attacked by Relevance theorists (Sperber and Wilson 1981, Wilson and Sperber 1992, Sperber and Wilson 1998, Wilson 2006). According to Sperber and Wilson, ironical utterances have four main properties: (i) They are mentioned, not used, (ii) they are echoic in nature, (iii) the ironical interpretation is an implicature that is derived through recognition of the echoic character of the utterance (Sperber and Wilson 1981: 309), (iv) the ironical speaker displays a dissociative attitude towards the proposition uttered. Take the utterance What a lovely weather! as an example. When uttered during a downpour, the speaker cannot mean the opposite because this would be uninformative. Instead they want to convey that it was absurd to assume that the weather would be nice. Thus, the ironical utterance is a case of an echoic mention of a previously entertained proposition. Types of echo include sarcastic repetition, attributed thoughts, norms and standard expectations (Sperber and Wilson 1998); thus, the echo theory of irony does not imply that there is always an original utterance that is exactly reproduced. (8)

(9)

(10)

A: B:

I’ll be ready at five at the latest. Sure, you’ll be ready at five.

[Sarcastic repetition]

A: B:

I’ll be ready at five at the latest. You mean at five tomorrow?

[Attributed thought]

A: B:

I’ll be ready at five at the latest. You are so punctual.

[Norm]

164   

(11)

   Interface Problems

A: B:

I’ll be ready at five at the latest. It’s a great virtue to be on time!

[Standard expectations]

According to Sperber and Wilson, the echo theory is constrained in that most utterances cannot be interpreted as echoes, and echoic interpretations must contribute to the relevance of an utterance. Several objections to this theory may be made (Lapp 1992; Sperber and Wilson 1998): (i) The notion of an echo is far too vague; it does not make sense to look for an echo in cases of conventional irony, e.g., uttering Boy, is it hot! when it is icy cold. (ii) Because not every echoic mention is ironical, echoic mention is not sufficient to explain ironical interpretation. (iii) It is not clear why the substitution by the opposite should not be a starting point in the search for the dissociative attitude of the speaker towards the proposition. (iv) Relevance theory cannot explain why hearers often fail to grasp the relevance of an ironical utterance. While there are cases in which echoing mention is typical for ironical remarks, the idea that echoing mention is involved in every ironical utterance appears far too radical. In her harsh attack of Sperber and Wilson’s theory, Roguska (2007b: 319) argues that the claim that every ironical utterance is connected with echoing a proposition renders the theory irrefutable. “In order to explain irony in the sense of S&W in such cases where it is impossible to find any original utterance that has been echoed, one has to assume that S herself produces some inadequate, irrelevant or absurd utterance, which she then mentions in order to express her dissociative attitude towards it. I want to emphasize that one has to make this assumption only in order to apply the notion of echo. Such an explanation of irony, however, seems to be very odd – to put it mildly.”

A competitor of the echo theory is the pretence theory. Recall that Grice (1989a: 34) argues that in ironical utterances, the speaker is “making as if to say” something; therefore, they may be seen as an adherent of the so-called pretence theory which views the ironic speaker as someone who pretends something (Grice 1989b: 54): “To be ironical is, among other things, to pretend (as the etymology suggests), and while one wants the pretence to be recognised as such, to announce it as a pretence would spoil the effect.” For instance, remarking What a lovely weather! during a downpour is pretending that the speaker finds the weather lovely, not asserting it. Wilson (2006: 1740) argues at length against pretence theories. She identifies a core set of ironical examples, and then goes on to claim that “pretence theories have to be supplemented with something like the notion of echoic use”. Moreover, according to her, “central cases of irony form a natural class with interpretive, attributive, and echoic utterances”. Since for her, pretence is nothing but “a

Irony and tautology   

   165

type of simulation or mimesis which crucially involves the exploitation of resemblances” (which is captured by the Relevance theoretic notion of echo anyhow), there is no need for a pretence theory of irony. The set of examples Wilson refers to is given in (12–18) below: (12)

[Mary (after a difficult meeting)]: That went well.

(13)

As I reached the bank at closing time, the bank clerk helpfully shut the door in my face.

(14)

Tim Henman is not the most charismatic tennis player in the world.

(15)

He is a fine friend. [see (6)]

(16)

Palmer gave Nicklaus quite a beating.

(17)

[Said to a neurotically cautious driver] a. Don’t forget to use your indicator.

(18)

b.

Do you think we should stop for petrol?

c.

I really appreciate cautious drivers.

Trust the Weather Bureau! See what lovely weather it is: rain, rain, rain.

I agree that in these examples, in an appropriate context, echoic mention may play some role (Wilson 2006: 1731–32); but I do not see why these examples should not be examples which show pretence. Lapp (1992) sees the disadvantages of both approaches, echoic as well as pretence theories, in their focus on the attitudes of speakers towards some proposition (i.e., the attitude towards an echoed or pretended proposition). His approach is embedded in a speech-act theoretical framework. The idea is to define irony as “Simulation der Unaufrichtigkeit” (simulation of insincerity), and hence to draw a distinction to lying that is “Simulation der Aufrichtigkeit” (simulation of sincerity): “Während die Lüge eine verdeckte Simulation der Aufrichtigkeit auf der Ebene der propositionalen Einstellung ist, ist die Ironie eine offene Simulation der Unaufrichtigkeit auf der Ebene sprachlicher Handlungen. Während der Lügner eine propositionale Einstellung simuliert, simuliert der Ironische eine sprachliche Handlung.“ (Lapp 1992: 146) [While the lie is a covert simulation of sincerity on the level of propositional attitude, irony is the overt simulation of insincerity on the level of verbal acts. While the liar simulates a propositional attitude, the ironical speaker simulates a verbal act. My translation, JM]

166   

   Interface Problems

What the ironical speaker simulates is that they are lying: “Der ironische Sprecher muss voraussetzen, daß der Hörer in der Lage ist, seine wahre propositionale Einstellung zu erkennen. Die Ironie ist also keine echte oder wirkliche Lüge, sondern eine simulierte Lüge.“ (Lapp 1982: 146) [The ironical speaker must presuppose that the hearer is able to detect their true propositional attitude. Hence, irony is not a true or real lie, but a simulated lie. My translation, JM] For example, in uttering He is a fine friend the speaker simulates that they believe that X is a fine friend while they do not actually believe that X is a fine friend. This “lie” is intended to be detected by the hearer. I think that this approach shows how a pretence theory can be worked out. That this theory draws on the insight that ironical utterances are speech acts, after all, is very convincing. As we have seen in a previous chapter, the notion of sincerity condition is somewhat problematic. However, it appears to be possible to replace it by the maxim of Quality. Thus, ironical utterances are pretences or simulations of a violation of the first maxim of Quality. In simulating the violation, the speaker is not violating the maxim but observing it. Let us now ask whether it is possible to lie by using ironical utterances. Whereas some researchers argue that it is impossible to lie with ironical utterances for the simple reason that ironical utterances are no assertions and are indicated by signals of irony (Chisholm and Feehan 1977: 152), others argue that this is indeed possible (Simpson 1992: 630). Of course, the basic problem is that in ironical utterances the implicature q may equal ‘not p’, as This was excellent is implicating ‘This was not excellent’, thus violating the logical principles of double negation, bivalency, and the excluded third. This was the main reason why Falkenberg (1982) refrained from the definition of ‘indirect lying’, as pointed out above. Look at the following example, drawn from Falkenberg (1982: 126); the example is based on Agatha Christie, Why didn’t they ask Evans?: (19)

Speaker S crashes into a wall in order to pretend that he had a traffic accident. An eyewitness H comes to the scene of accident and cries: “Jesus! Did you have a traffic accident?” Speaker S replies: “No. I crashed into the wall deliberately.”

The opposite of S’s utterance, ‘I didn’t do it deliberately’, certainly is a good candidate for an ironical implicature. In this case, the proposition p that S crashed into the wall deliberately is true, while the implicature q that they didn’t do it deliberately is false. An echoic analysis seems also plausible. What is echoed is the standard expectation that no one crashes into a wall deliberately. Thus, the speaker dis-

Irony and tautology   

   167

sociates themself from the possibility that someone crashes into the wall deliberately. In this special context, however, they lied. Note that the case of lying while being ironical forces us to analyze both p and q. Consider the ironical utterance This was clever implicating ‘This was not clever’. If the implicature is untruthful because the speaker believes that it was clever, then we say that they were lying while being ironical. They did not want to utter directly and unironically This was clever in order to deceive the hearer. The same holds for the reverse situation when the speaker utters This was not clever implicating ‘This was clever’. Let us now go into the question whether lying by using tautologies is possible. Tautologies have been repeatedly discussed (see, among others, Levinson 1983; Fraser 1988; Wierzbicka 1987; Ward and Hirschberg 1990; Autenrieth 1997; Davis 1998; Meibauer 2008; Rhodes 2009), the basic observation being that a typical tautology of the form ‘Ni is Ni’, e.g., Business is business, is always true. Therefore, its utterance seems to be uninformative, and the speaker is suspected to violate the first maxim of Quantity saying that the speaker should make their contribution as informative as is required (for the current purposes of exchange). However, the speaker may be considered cooperative if the hearer assumes that the speaker wants to implicate ‘There is nothing one can do about it’ (or the like). This implicature may be false from the point of view of the speaker, and thus is a candidate for an extended definition of lying that includes falsely implicating. The case of tautological utterances is particularly interesting because of the fact that the speaker cannot lie while uttering p. When using a tautology, lying is only possible through falsely implicating. Let us discuss whether uttering a tautology is asserting, whether it is possible to lie by using p, and whether it is possible to lie by using q. First, note that Levinson (1983: 110) assumes that tautologies are assertable. This is in line with our definition of assertion: ‘Ni is Ni’ is a declarative sentence with the meaning p, p is represented as true by the speaker, and the speaker M-intended that an addressee actively believes that p. If someone denies that tautologies may be asserted, then this is because of the idea that something has to be clearly contingent to qualify as an assertion. However, I see no problem in assuming that tautologies are asserted by necessity in order to convey the implicature that is bound to them. Second, when uttering a tautology, it seems not to be possible to actively believe that p is false. Therefore, it is not possible to properly lie by using a tautology, and we would have a violation of the condition that the speaker does not believe p in our definition of lie. If the speaker utters War is war, and the hearer objects This is not true, then the hearer objects to the implicature (e.g., ‘There are no good wars and bad wars’, ‘There is nothing one can do about it’, etc.), not to the propositional content.

168   

   Interface Problems

Third, falsely implicating while using tautologies is no problem. Just imagine a pacifist (spying out a group of rednecks) uttering War is war, thereby falsely implicating ‘There is nothing one can do about it.’ In our approach, it is possible to express the idea that tautologies may be lies: p is necessarily true, while q may be true or false. While most approaches, if not sceptical about a Gricean approach to implicatures altogether (Wierzbicka 1987; Davis 1998), assume that the maxims of Quantity and/or Relevance are involved, Autenrieth (1997) argues that, at least in standard examples like Bachelors are bachelors, the sequence are bachelors is understood as a predication. In this approach, the predicationally used bachelors gives rise to conceptual enrichments that figure prominently in the derivation of the associated implicature that is triggered by the maxim of Relevance. Thus, an utterance like Bachelors are bachelors translates into something like ‘All bachelors possess the typical properties of bachelors that are mutual knowledge in the speaker community.’ If this is not a logical tautology, as Autenrieth argues, then it may be asked whether it is possible to lie by such an utterance (note that we focus on the assertion here). Consider the following exchange: (20)

B: A:

Pete has not washed his socks for months. Bachelors are bachelors. [+> ‘There is nothing one can do about it.’]

Let it be part of the literal meaning of bachelor that bachelors are unmarried, and part of the stereotype of bachelor that they are not neat. If speaker S falsely believes that some bachelors are married, then it is not possible to take their utterance as a lie; it is simply not possible to have a corresponding M-intention. If, on the other hand, the speaker thinks that bachelors are neat, I think that their utterance would not count as a lie either because S’s conviction is not shared with the speaker community. Therefore, I still think that it is not possible to lie by using mere tautologies. The point is that the primary job of tautologies is to convey implicatures, not to assert propositions. And it is by intentionally false implicatures that lying is possible, as has been proposed in chapter 4.

5.5 Understatement and overstatement (hyperbole) In everyday language, overstatements and, to a lesser degree, understatements are quite common. Yet, there is not much research on this phenomenon (but see Nemesi 2010, Claridge 2011). Overstatements, such as the utterance I have tried to call you a thousand times!, spoken in a reproachful tone, have been analyzed as “partial lies” with “reduced truth value” (Fónagy 2001: 218). If the speaker is

Understatement and overstatement (hyperbole)   

   169

committed to the truth of their utterance, i.e., to the truth of the proposition ‘S has called H a thousand times’, and this is false, then one could indeed conclude that S is lying. However, overstatements are statements (assertions), yet are usually not considered as lies. Instead, they are merely deliberately inexact representations of a state of affairs. Drawing on examples taken from Colston and Keller (1998), Nemesi (2010) presents the following scenarios with overstatements and understatements: Suppose that Ken and Barbie stay in an apartment that was left to them for a short vacation by a close friend. They open the fridge and see that it is full of food (Context, 1). (21)

a.

This fridge has tons of food in it. [positive overstatement]

(‘It is full of food.’)

b.

This fridge has some food in it. [(ironic) negative understatement]

(‘It is full of food.’)

When Ken utters (21a) it is evident to both interlocutors that this is not the truth: the fridge is full, but there are not literally “tons” of food in it. When Ken utters (21b), he states that there is food but not much, so he downplays the amount of food. In both cases, he deviates from the truth. We call (21a) a positive understatement and (21b) a negative understatement. Now suppose another context (Context, 2) in which the fridge is not totally empty, but the food in it is not enough. (22)

a.

This fridge has nothing in it. [negative overstatement]

(‘It is nearly empty.’)

b.

This fridge has quite a lot of food in it. [(ironic) positive understatement]

(‘It is nearly empty.’)

Then Ken negatively overstates with his utterance in (22a), while with (22b), he positively understates. It becomes clear that scales of amplification/augmentation and attenuation/diminution are involved, and that the hearer has to find out what the adequate scale is. (The qualifications of (21b) as “negative” and of (22b) as “positive” are added by me. Why Nemesi speaks of “ironic” understatements here is not clear to me; I suspect that this has to do with the fact that an interpretation of “the contrary” comes into play.) Nemesi (2010: 384) states that “in case of each over- and understatement it is supposed that the speaker could have used a more adequate linguistic expression referring to the real state of affairs, but for some reason or goal, s/he wanted to express something more (his/her expectation, impression, attitude, mood, etc.)” (Nemesi 2010: 384). It comes as

170   

   Interface Problems

no surprise, then, that the additional meanings have the status of conversational implicatures. The question is, which maxims are involved. There are three possibilities: First, the maxim of Quality, because the speaker is not truthful. This is the classic view of Grice, but recall that his category of Quality-violating tropes, including irony, sarcasm, metaphor, and meiosis, is far too wide to be convincing without further qualification. The second is the maxim of Quantity, insofar as, e.g., overstatement is conceived of as sort of too much information, and understatement as not enough information. I think that some types of exaggeration where too much detail is given are correctly described with respect to Quantity, but the cases (21) and (22) do not belong into this category. So Nemesi (2010: 401) is right to rebut this view, arguing that “it is no more (or less) informative than is necessary but false or inaccurate information” – this in turn is a plea for the maxim of Quality. The third possibility not considered by Nemesi is the maxim of Manner, since it is not perspicuous, e.g., to utter This fridge has tons of food in it when what is really meant is ‘It is full of food.’ However, as Levinson (2000) makes clear, the M-principle has its merits when two alternative ways of saying something are synonymous, i.e., have the same denotation. And this may be debated in the case of tropes like overstatement, ironies, and metaphors. Thus, the maxim of Quality is still the best candidate for a maxim playing a role in the derivation of conversational implicatures related to over- and understatement. The surplus value of overstatements is revealed by the concrete inference pattern for overstatements proposed by Nemesi (2010: 399) because here he assumes three types of suppositions (implicatures) connected to the overstatement p: […] first, supposition q is needed, which is the adequate transformation of p on the relevant semantic scale of amplification or attenuation, concerning its relation to the real state of affairs […]; second, supposition r as a “weak” implicatum is needed for recovering the speaker’s attitude (his/her emotion, evaluation) towards the topic of conversation on the basis of the hyperbole expressed (r can also be conventionalized, at least partly; however, the stylistic value of the expression holds the subjective overtones); and third, insofar as the context and supposition r jointly support such an additive inference, supposition s as another weak implicature is needed as well, which reveals the selfpresentational role of p manifested through r; […].

The given context (in fact, taken from an Hungarian classic movie) is that Gitta answers to Kalocsay who supposes that he needs one and a half years to win her

Lying and bullshitting   

   171

heart: You would not succeed even within a thousand years. According to Nemesi (2010: 407–408), this overstatement is connected with implicatures q, r, and s: (23)

p = You would not succeed even within a thousand years. q (strong implicature) = Kalocsay could not or would not conquer Gitta. r (weak attitude implicature) = Gitta categorically refuses the idea of Kalocsays’s courting. s (weak self-implicature) = Gitta is a moral young lady, who is not so easy to conquer as Kalocsay thinks.

Now imagine a situation in which Gitta secretly loves Kalocsay. Then, according to my proposal that untruthful implicatures are a part of the lying act, her utterance would count as a lie. And indeed, to come back to our introductory example: If I say the utterance I have tried to call you a thousand times!, thereby implicating that I tried to call you, say, six times, while in fact I only tried it two times, I would call this a lie. Hence, overstatements and understatements seem well-suited for lying; because of their emotional overtones and their explicit way of saying something which is not the truth, they invite lying, especially where the real state of affairs, as in the fridge case, is not evident for the participants. The crucial point is, then, that over- and understatements may hide truthful evaluations of states of affairs.

5.6 Lying and bullshitting Harry G. Frankfurt, in his seminal essay “On Bullshit” (2005), assumes that there is a sort of language activity act named bullshit. (This essay was published for the first time in the journal “The Raritan Review”, VI:2, 1986, then in a collection of Frankfurt’s essays, “The Importance of What We Care about: Philosophical Essays”, 1988). According to Frankfurt, the producer of bullshit is the bullshitter, the act of producing bullshit is bullshitting. Hence, it appears that bullshit is a type of speech act, and thus a pragmatic category. On the first page of his essay, Frankfurt appeals to our intuition that there is a cultural phenomenon called bullshit: “One of the most salient features of our culture is that there is so much bullshit. Everyone knows this. Each of us contributes his share. But we take the situation for granted. Most people are rather confident of their ability to recognize bullshit and to avoid being taken in by it. So the phenomenon has not aroused much deliberate concern, nor attracted much sustained inquiry.” (Frankfurt 2005: 1)

172   

   Interface Problems

However, according to Frankfurt, there is no theory of bullshit. “In consequence, we have no clear understanding of what bullshit is, why there is so much of it, or what functions it serves. And we lack a conscientiously developed appreciation of what it means to us. In other words, we have no theory.” (Frankfurt 2005: 1)

In his essay, Frankfurt famously succeeds in showing us important aspects of the phenomenon, yet he does not present a clear definition of what bullshit is. In this section, I will focus on the question whether bullshit is a reasonable pragmatic category. If bullshit is a kind of speech act, it should be analyzable along the lines of speech-act theory, i.e., it should be possible to define bullshit through felicity conditions. However, Frankfurt does not give any felicity conditions while offering, in a largely inductive manner, aspects of a definition of bullshit. First of all, we need to have a convincing example for bullshit. Unfortunately, the only example Frankfurt gives is not convincing. This example reports an exchange between Fania Pascal and Ludwig Wittgenstein. Fania Pascal writes: “I had my tonsils out and was in the Evelyn Nursing Home feeling sorry for myself. Wittgenstein called. I croaked: ‘I feel just like a dog that has been run over.’ He was disgusted: ‘You don’t know what a dog that has been run over feels like.’” (Pascal 1984, quoted after Frankfurt 2005: 24)

Our first impression is that Wittgenstein took Pascal’s utterance too seriously, and that he thus ignored that it was meant as a jokey simile. Yet Frankfurt argues: “The trouble with her statement is that it purports to convey something more than simply that she feels bad. Her characterization of her feeling is too specific; it is excessively particular.” (Frankfurt 2005: 29)

However, this seems to be a matter of taste. Pascal’s utterance has nothing to do with knowledge about the feelings of run-over dogs, and how one can be certain about these feelings. As we will see in a minute, Frankfurt argues that “bullshit” is characterized by a “lack of concern for the truth” and a “misrepresentational intent” on the part of the speaker – however, both ingredients are missing in the example of the unlucky dog (Maes and Schaubroeck 2006). I will present two examples that – intuitively – are better suited to illustrate bullshitting. The first is an advertisement for “Fractional Neck Lift Concentrate” that I found in the magazine “Wallpaper” from 2009. For me, it is an impressive piece of advertising bullshit.

Lying and bullshitting   

(24)

   173

Fractional Neck Lift Concentrate The first in fractionated topicals. With discovered-on-Mars Iron Rose Chrystals. Y-42 FRACTIONAL NECK LIFT CONCENTRATE More than fractional treatments. For less than fractional treatments. In simple terms: neck therapy at its avant-garde finest. Protein fractions maximize fibrillin synthesis and minimize the inevitable, irreversible degradation of elastic tissue. Discovered-on-Mars iron rose crystal comes from effusive magma rock to increase prolyl hydroxylase activity by 381% and boost collagen production. Our new tetrapeptide-9 from France, as well as a next-generation tripeptide-10 citrulline and a bi-blinked dipeptide from Switzerland, to stimulate Laminin V, Collagen IV, Collagen VII, Collagen XVII and Integrin. A new tetrapeptide-11 from France also comes to the rescue to stimulate Syndecan-1 synthesis and reinforce epidermal cohesion. From Mother Nature, clinically-affirmed biotechnological derivatives of European elder, sweet chestnut and shiitake not only firm the skin, but also improve its protective responses dramatically. Finally, controlled organism ferments from Spain and Switzerland offer tightening of the skin’s appearance instantly and in the long term. And the active clinical results? Too many to name, but a 350% improvement in wrinkle appearance during an 84-day third-party study of 60 volunteers stands out as unprecedented proof. [Avant-garde aroma: earth tones and floral roots] [Wallpaper, October 2009]

Intuitively, the whole text strikes me as overwhelmingly unclear. In order to focus on a speech-act analysis of bullshitting, I would like to draw the reader’s attention to the last lines: (25)

And the active clinical results? Too many to name, but a 350% improvement in wrinkle appearance during an 84-day third-party study of 60  volunteers stands out as unprecedented proof. [Avant-garde aroma: earth tones and floral roots]

In particular, the assertion “a 350% improvement in wrinkle appearance during an 84-day third-party study of 60 volunteers stands out as unprecedented proof” strikes me, intuitively, as being bullshit. The second example is a press release concerning the separation of the German model Heidi Klum and singer Seal:

174   

(26)

   Interface Problems

Separation of Heidi Klum and Seal In a joint statement to E! News, the couple said, “While we have enjoyed seven very loving, loyal, and happy years of marriage, after much soulsearching we have decided to separate. We have had the deepest respect for one another throughout our relationship and continue to love each other very much, but we have grown apart. This is an amicable process and protecting the well-being of our children remains our top priority, especially during this time of transition. We thank our family, friends, and fans for their kind words of support. And for our children’s sake, we appreciate you respecting our privacy.“ [http://prweb.com/releases/actingauditions/modelingauditions/ prweb9142093.htm] [01/27/2012]

Here, my intuition is that the utterance (27) is a good example of bullshit: (27)

While we have enjoyed seven very loving, loyal, and happy years of marriage, after much soul-searching we have decided to separate.

Before we come back to these texts, we have to delimit the putative pragmatic category bullshit against its neighbors – that is, against humbug, hyperbole, and lying. A definition of humbug, actually one that inspired Frankfurt’s work, was given by Max Black in his essay “The Prevalence of Humbug”. According to him, humbug is a “deceptive misrepresentation, short of lying, especially by pretentious word or deed, of somebody’s own thoughts, feelings, or attitudes” (Black 1983: 143). A typical example is, as Frankfurt points out, the Fourth of July orator who “goes on bombastically about ‘our great and blessed country, whose Founding Fathers under divine guidance created a new beginning for mankind’” (Frankfurt 2005: 16). While bullshit has to do with misrepresentation and a deceptive intent as well, pretention or relatedness to subjective thoughts, feelings, and attitudes are not necessary ingredients of bullshitting. Next, consider hyperbole. Hyperboles like I told you a thousand times to shut the door! may be defined as deliberately inexact representations of a state of affairs. They are literally false, and the speaker shows no interest in describing states of affairs in an exact way. However, there is no intent to deceive involved. The speaker wants the hearer to derive a certain conversational implicature, like in irony. In contrast, bullshit seems to be somehow related to deception. This leads to our third and most important issue, lying. Recall the definition of lying developed in chapter 4. If the speaker asserts p and thereby represents p as true while in fact they believe that not-p, then they are lying. While “bullshit”

Lying and bullshitting   

   175

is also connected to the intent to deceive (which follows from the insincere assertion), the bullshitter may (to a certain extent) believe that p. Frankfurt’s “bullshit” heuristics basically comprises two ingredients. The first ingredient is a certain “lack of concern for the truth”: “The fact about himself that the bullshitter hides, on the other hand, is that the truth-values of his statements are of no central interest to him; what we are not to understand is that his intention is neither to report the truth nor to conceal it.” (Frankfurt 2005: 55)

The second ingredient is the bullshitter’s “misrepresentational intent”, also called “phoniness”: “Since bullshit need not to be false, it differs from lies in its misrepresentational intent. The bullshitter may not deceive us, or even intend to do so, either about the facts or about what he takes the facts to be. What he does necessarily attempt to deceive us about is his enterprise. His only indispensably distinctive characteristic is that in a certain way he misrepresents what he is up to.” (Frankfurt 2005: 54)

In the following, I will stick to Frankfurt’s intuitions and propose a definition of bullshitting that is based on the idea that bullshitting is an insincere act of asserting. So let us first have a look at my definition of assertion, here presented in a simplified version: (28)

Asserting The speaker asserted at time t that p iff (a) the speaker uttered a declarative sentence meaning p, (b) the speaker presented p as true, (c) the speaker intended that the addressee believed p.

For instance, the utterance in (25), A  350% improvement in wrinkle appearance during an 84-day third-party study of 60 volunteers stands out as unprecedented proof counts as an assertion. We can then go on and define bullshitting in the following manner: (29)

Bullshitting (reconstructing Frankfurt) The speaker was bullshitting at time t that p iff (a) The speaker acted as if he asserted that p, however: (b) [loose concern for the truth] the speaker neither intended to present p as true nor intended to present p as not true.

176   

   Interface Problems

(c) [misrepresentational intent] the speaker intended that the addressee did not become aware of (b). Thus, while in a sincere assertion, the speaker presents p as true, in bullshitting the speaker is ignorant with respect to matters of truth and falsity. Carson (2010: 58) argues that the following three assumptions of Frankfurt are wrong: (a) bullshitting requires an intention to deceive; (b) bullshitting is not lying; (c) bullshitters, as opposed to liars, are indifferent to truth. All of his counterexamples are related to so-called evasive bullshitting, i.e., bullshitting in situations of social pressure (see also Coleman 2007 on evasion). The example of the lazy student is directed against (a). A lazy student produces a bullshit answer to an exam question. She knows that it is better to offer something instead of nothing, knowing that her answer is bullshit and assuming that her professor knows that, too. However, I maintain that the student’s answer is still assertive, hence condition (29a) remains intact; note that the student avoided truthfully writing I don’t know the correct answer to the question or the like. As an example directed against (b), Carson presents a scenario in which a politician who is under pressure produces bullshit and lies at the same time. Similarly, with respect to (c), Carson argues that evasive bullshitting is compatible with a concern for the truth. I do not find these cases convincing, because mixing bullshitting and lying does not amount to being unable to distinguish the two. Carson (2010: 62) ends his attack on Frankfurt’s concept of bullshitting with the following confession: “I do not have a better alternative definition of bullshit that I am prepared to defend. I suspect that the concept of bullshit is too loose and amorphous to admit of a definition in terms of necessary and sufficient conditions.” In contrast, I think that (29) provides a rational basis for clarifying the concept. In the following, I propose a further condition that is also relevant to Carson’s scenarios of evasive bullshit (see also Meibauer 2013b). My intuition is that certainty and uncertainty play an important role in bullshit, a role that has hitherto been neglected. For instance, the utterance A  350% improvement in wrinkle appearance during an 84-day third-party study of 60 volunteers stands out as unprecedented proof is presented with outstanding certainty on the part of the writer. It is backed by scientific authority as the whole text in (24) demonstrates, a text that is nearly incomprehensible for the average reader. There are, of course, numerous definitions of certainty in epistemology. I stick to a classical definition of (subjective) certainty proposed by Roderick Chisholm (see also section 7.5):

Lying and bullshitting   

(30)

   177

Certainty (Chisholm 1976: 27) p is certain for S at t =df (i) Accepting p is more reasonable for S at t than withholding p (i.e., not accepting p and not accepting not-p) and (ii) there is no q such that accepting q is more reasonable for S at t than accepting p.

The intuition that the bullshitter expresses too much certainty with respect to p might have to do with a certainty norm for assertion occasionally touched upon in the literature: (31)

Certainty Norm for Assertion Assert that p only if you are certain that p. (Unger 1975, Stanley 2008)

Now it appears that in bullshit, this norm is systematically broken. For instance, Heidi Klum may not be certain that she “enjoyed seven very loving, loyal and happy years of marriage” and that she engaged in “much soul-searching”; nevertheless she is asserting this with great certainty. Withholding p and accepting q are no option for her at all. Moreover, there seem to be degrees of certainty, that is, certainty is a scalar notion: (32)

Degrees of Certainty The speaker is at t certain that p to a high degree, iff (i) the addition of q to their knowledge would not make them more certain, and (ii) the addition of r to their knowledge would not make them less certain.

So the speaker is in a stable position of certainty with respect to p. On the other hand, there is a possibility of becoming more or less certain if the knowledge base is changed. Suppose that Ken asserts that Barbie has a Ph.D. (= p). Now Ben tells Ken that he saw a doctoral cap in Barbie’s apartment (= q). But this does not make Ken more certain. Although his assumption is corroborated by Ben’s information, he does not change his mind. This would only happen, if he was in doubt regarding p.

178   

   Interface Problems

If Ken, however, asserted that Barbie had not gone to college (= r), Ken could become less certain about his belief. He knows that having visited college is a precondition for having a Ph.D. Now Ken is very certain about his knowledge (he is stubborn). He thinks that even if the information given by Ben is true, there must be a possibility to do a doctor’s degree without having visited a college. So Ken blocks his becoming less certain. Hence, the revised definition of bullshitting contains a new condition (d) that refers to “too much certainty” on the side of the bullshitter. (33)

Bullshitting (revised) The speaker was bullshitting at time t that p iff: (a) The speaker acted as if they asserted that p, however: (b) [loose concern for the truth] the speaker neither intended to present p as true nor intended to present p as not true. (c) [misrepresentational intent] the speaker intended that the addressee does not become aware of (b). (d) [too much certainty] the speaker presented p with far more certainty than would be adequate considering their loose concern for the truth (see (b)).

So bullshitting is a speech act that masks itself as a proper assertion, yet is deficient as a proper assertion in a number of dimensions. Whether (d) is necessary as a separate condition is of course debatable. As Christian Plunze (p.c.) points out, condition (33d) may be derived from the certainty norm for assertions in connection with condition (33b). When bullshitting is pretence of assertion, it is implied that the speaker obeys to the certainty norm. However, since the speaker has a loose concern for the truth, they cannot but violate the certainty norm. I agree, in principle; yet I see an advantage of condition (33d) in that it pinpoints the stark discrepancy between the apparent assertion and the hidden loose concern for the truth. Let us now see whether this definition applies to the case of Heidi Klum and Seal’s press release: (a) The authors assert that they “enjoyed seven very loving, loyal and happy years of marriage” (p1) and that they “continue to love each other very much” (p2). Normally, one would not infer from p1 and p2 the necessity of a separation. (b) The authors neither want to present p1, p2 as true nor to present p1, p2 as not true. How p1, p2 are related to the “growing apart” of the authors and

Lying and bullshitting   

   179

motivate the separation remains unclear. Quite possibly, it is unclear to the authors themselves. (c) The authors want the addressee to not become aware of the fact that they present p1, p2 neither as true (although this is what a hearer may conclude on the basis of their assertion) nor as not true (the latter being compatible with the separation). (d) The authors lack the necessary certainty for the assertion of p1, p2 (otherwise they would not separate), yet assert p1, p2 with certainty. It goes without saying that both texts are persuasive in character. From the perspective of persuasion, it can be concluded that bullshitting may have the effect to render a reader basically to trust the speaker more certain or more uncertain. Suppose that, initially, Ken does not believe that it is possible to get a 350% improvement in wrinkle appearance. However, if he comes to hear from the clinical study (by reading the advertisement), he is getting almost certain of this possibility. Certainty has to do with potential justification. Ken is certain insofar as he now assumes that (i) the addition of another proposition q to his knowledge would not make him more certain, and (ii) the addition of another proposition r to his knowledge would not make him less certain. If, on the other hand, Ken is doubtful with respect to p or detects that the speaker asserted p with more certainty than would be adequate (on the basis of the speaker’s putative loose concern for the truth), then this could render Ken more uncertain. For example, while he always had confidence in the relationship of Heidi Klum and Seal, after reading their press release his confidence is shaken to the core. In fact, he can even imagine that there might be further information that would render him still more uncertain. In his strict verdict of bullshit, Frankfurt goes so far as to state: “He [the bullshitter, JM] does not reject the authority of truth, as the liar does, and oppose himself to it. He pays no attention to it at all. By virtue of this, bullshit is a greater enemy of the truth than lies are.” (Frankfurt 2005: 61).

But then the question is why we are so tolerant of bullshit, as compared to lies. This could have to do with (i) tolerance of the general goals of speakers (for instance, to sell Fractional Neck Lift Concentrate or to avoid loss of image), and (ii) funny effects triggered by bullshitting (see also bull session, bullshit bingo, etc.). The latter effect may be explained by the stark discrepancy between the claimed certainty of the speaker vis-á-vis the lacking substance of the utterance’s content.

180   

   Interface Problems

In the remainder of this chapter, let us have a look on the taxonomy of bullshit texts and on the proposal by McGinn (2008) to establish a deceptive triad including lying, bullshitting, and mindfucking. In my analysis, I concentrated on bullshitty texts, singling out portions of it that were promising candidates for the status of bullshitting. In fact, I assume that there are single utterances constituting bullshit as well as complete texts. Every bullshit text should contain at least one speech act of bullshitting. Hence, we can ask whether there are typical textual properties of bullshit. A proper answer of course demands more empirical research into bullshit; however, we have initial evidence that this is the case. Have a look at (24) again. As far as coherence is concerned, a load of bullshit comes with the enumeration of locations from which the ingredients for the Fractional Lift Concentrate are taken, namely from the Mars, France, Spain, Switzerland, and Mother Nature. With regard to cohesion, there obviously is a lack of it in the sequence at the end of the text when we read: Too many to name, but a 350% improvement in wrinkle appearance during an 84-day third-party study of 60 volunteers stands out as unprecedented proof. [Avant-garde aroma: earth tones and floral roots]. Here, it is unclear how the assertion relates to the information in brackets. Another aspect of bullshit texts is their relation to different domains such as politics, advertising, and the media. Whereas Frankfurt (2005) is interested in a general definition of bullshit, Cohen (2002) is solely interested in academic bullshit. According to Cohen (2002), an essential feature of pieces of academic bullshit is their unclarifiability (Cohen 2002: 332): (34)

Unclarifiability Something is unclarifiable if and only if it cannot be made clear (…).

Cohen (2002: 333) refrains from defining what clarity is but offers a “sufficient condition of unclarity”: (35)

Negation test for unclear texts “(…) adding or subtracting (if it has one) a negation sign from a text makes no difference to its level of plausibility.”

The problem with this test is, of course, that it is not clear how “plausibility” is to be defined. There may be a huge amount of variation between speakers when they are asked whether a certain text sounds plausible to them. It is this reluctance to clarify the notion “clear” that Frankfurt (2002) attacks in his reply to Cohen. Moreover, he proposes a translation into clear English of a putative piece

Lying and bullshitting   

   181

of academic bullshit that, according to Cohen’s analysis, was unclarifiable by definition. (36)

Putative academic bullshit This is precisely the first meaning we can give to the idea of dialectic: a logic or form of explanation specifically adapted to the determinant intervention of class struggle in the very fabric of history. (Etienne Balibar)

(37)

Frankfurt’s translation The most distinctive point of dialectical explanations is precisely that they are particularly helpful in illuminating how class struggle has significantly determined the course of history.

This demonstrates to a certain extent that it is possible to make an unclear text clearer. What the linguist is interested in, however, are the verbal indicators that make a text unclear (see also the taxonomy proposed in Evans 2006: 198–199). We will come back to the issue of clarity in section 7.4. McGinn (2008) views mindfucking in relation to lying on the one hand, and bullshitting on the other hand. They belong to one class because they have to do with an abuse of truth. Mindfucking is defined as “some sort of assault on the mind, an invasive operation performed on the psychological state of the person” (McGinn 2008: 1). Taking up the metaphor of “fucking”, he ventures that there are positive and negative variants of this act (McGinn 2008: 5). His prime interest is in the negative variant, and he stresses that negative mindfucking always aims at triggering a bad feeling in the addressee (McGinn 2008: 36): “[…] the point I am making is that it is not intrinsic to bullshitting that it aims at psychological disturbance of an emotional kind, whereas this is intrinsic to mindfucking.“ This negative feeling can have the property of lasting harm (McGinn 2008: 41). In sum, then, mindfucking is an “uninvited and malevolent manipulation of someone’s mind and beliefs through illegitimate, non-rational, emotion-based techniques”, as Dov Lerner (2009: 123) puts it. Although the comparison to lying and bullshitting seems to suggest a close relationship, mindfucking is, in my view, not a type of speech act in the same way as bullshitting is. All I can think of is the utterance of ethnic slurs, but this is not an example provided by McGinn. Thus, the value of the concept of mindfucking is more of a metaphorical nature and amounts to well-known concepts like brainwashing, propaganda, etc. In summary, then, the initial question whether bullshit is a reasonable pragmatic category can be answered positively. Drawing on insights of Harry Frankfurt’s seminal essay On Bullshit (2005), bullshit is defined as an act of insincere

182   

   Interface Problems

asserting in which the speaker shows (a) a loose concern for the truth, (b) does not want the addressee to become aware of condition (a), and (c) expresses more certainty than is adequate with respect to condition (a). This approach has been illustrated with two examples of putative pieces of bullshit. Potential counterarguments against the spirit of Frankfurt’s approach ventured by Carson (2010) are refuted. It has been pointed out that condition (c) may have to do with humorous effects of bullshit. Furthermore, it has been argued that McGinn’s (2008) notion of mindfucking does not lend itself easily to a speech-act analysis in analogy to bullshitting.

5.7 Conclusions In this chapter, we have discussed a number of thorny issues related to deception and lying, yet not easily located in minimalist or contextualist models of the semantics-pragmatics-interface. If truth-conditions were aspect-sensitive, then this would amount to another case of pragmatic intrusion into truth-conditional representations since the conditions of uses of certain names are certainly pragmatically determined. Fiction, so it seems, opens up another world that is accessible from our natural world, yet is not submitted to usual judgments of truth and falsity. Fiction is not lying, but there are lying fictional characters, and even authors may be liars when they are talking in the first person. Irony is another case that has to do with a second level of meaning that must be amenable to the hearer. While ironies are not lies, it is nevertheless possible to lie while being ironic. In addition, hyperboles are cases in which the speaker is not lying, but nevertheless they are not saying the truth. Finally, we have been going into the definition of bullshit, as famously portrayed by Harry Frankfurt. Bullshit comes close to lying in that it disrespects the truth, yet the bullshitter’s primary aim is not to deceive the hearer. The task is, then, to integrate these aspects in a more comprehensive picture of the semantics-pragmatics interface.

6 Lying and quotation 6.1 Introduction The reproach to have quoted something falsely is quite common, not only with respect to the media but also to everyday communication. This shows that there is a common presumption that it is possible to quote correctly, and moreover that one should, in the interest of truth and truthfulness, quote correctly. Even selective quoting may be regarded as deceptive in some cases (McGlone 2009). In the academia, certain standards of quotation have evolved, and these standards establish a certain norm with respect to correct quotation. For instance, a common practice of academic deception is to borrow ideas or texts from other authors without marking this. “Plagiarism” led to the resignation of the former German Minister of Defence, Karl Theodor Freiherr zu Guttenberg, in 2010. In this chapter, we will explore the relation between intentionally false quotations and lying (see Meibauer 2007). Recall the definition of assertion developed in section 3.11, repeated here for convenience: (1)

Assertion S asserted at t that p, if and only if (a) S uttered at t the declarative sentence σ meaning p, (b) by uttering the declarative sentence σ, S presented p as true, (c) by uttering the declarative sentence σ, S M-intended that an addressee H to whom S uttered p actively believes that p.

The act of quoting usually requires an original utterance (OU) that is somehow represented. This is illustrated in (2): (2)

a.

The project is hard to understand. [original utterance by Ken]

b.

Ken said:

“The project is hard to understand.” [quotation]

c.

Ken said:

“The project is easy to understand.”

d.

Ben said:

“The project is hard to understand.”

If the original speaker Ken (= OS) uttered (2a) at a certain point in time, another person may report his utterance by using (2b), a direct quotation. This leads us to the following definition of quotation (note that quotation refers to the whole

184   

   Lying and quotation

act of quoting, not only to the quoted material, i.e., the words within quotation marks): (3)

Quotation S quoted at time t by asserting that p iff (a) S uttered at t the declarative sentence σ meaning p, (b) p contained the representation of another sentence or sentence fragment υ meaning r, (c) this sentence or sentence fragment υ meaning r had been uttered by OS at time s< t, (d) by uttering the declarative sentence σ, S presented p as true, (e) by uttering the declarative sentence σ, S m-intended that an addressee H actively believed that p.

Now it is clear that is possible to lie in the act of quoting. If S quotes OS by uttering (2c) while believing that the original utterance was (2a), then this counts as a lie. The same holds for (2d) because S asserted that the OS was Ben, knowing that in reality it was Ken. In section 4.2, lying was defined as in (4), here repeated for convenience: (4)

Lying S lied at t, if and only if (a) S asserted at t that p, (b) S actively believed at t that not p.

In sum, then, quoting is a type of assertion, and since we analyzed lying as a type of assertion as well, untruthful quoting counts as lying, too. Let us turn to the taxonomy of quotations. In (3b), we mentioned “representation of another sentence or sentence fragment υ meaning r”; this is intended to cover the following types of quotation (cf. Brendel, Meibauer, and Steinbach 2011):

Lying, direct/indirect quotation, and the verbatim assumption   

(5)

   185

Types of Quotations Be the OU of Ken (OS) The project is hard to understand. Direct Quotation:

(a) Ken said: “The project is hard to understand.” Indirect Quotation: (b) Ken said that the project is hard to understand. Parenthetical Quotation: (c) The project, said Ken, is hard to understand. Mixed Quotation: (d) Ken said that the project is “hard to understand”. Scare Quote: (e) Ken said that the “project” is hard to understand. Free (in)direct Quotation: (f) Ken accelerated his steps. The project is hard to understand. But he would manage it. Emphatic Quotation: (g) “project” [written on a door] Pure Quotation:

(h) “Tcejopr” is not an English word.

Note that free, (in)direct quotations, emphatic quotations, and pure quotations are not quotations in the sense of definition (3) since they are not related to an original utterance OU. Scare quotes may or may not represent an original utterance. In the following, I will concentrate on direct quotation, indirect quotation, mixed quotation, and scare quotes. In the first place, we will discuss under which conditions a quotation will count as a lie, and whether the different types of quotation are related to a different lying potential.

6.2 Lying, direct/indirect quotation, and the verbatim assumption The essential properties of direct and indirect quotations can be sketched as follows. In prototypical direct quotations, (a) the original utterance (OU) is represented without any formal change (“verbatim”), (b) the represented OU is posited syntactically unconnected before or after a sentence containing a verb of saying (verbum dicendi), (c) first person deictic elements in the OU are related to the speaker of the OU, cf. (6b). The verb to say (Germ. sagen) counts as the prototypical verb of saying (cf. Partee 1973: 410).

186   

   Lying and quotation

In prototypical indirect quotations, (a) the original utterance appears as the complement of a verb of saying, (b) first-person deictic elements contained in the original utterance are shifted (6c) while first-person deictic elements not contained in the original utterance are related to the speaker of the quotation. Thus, if (6d) is spoken by Barbie, Ken said that Barbie is cool. Note that languages behave differently with respect to the formal mechanisms indicating quotations. For instance, in German the subjunctive may be used to indicate indirect quotation, and besides verb-final position forced by a complementizer (dass engl. that) verb-second (without any complementizer, e.g. in Ken sagt, Ø er ist cool) is also possible. (6)

a.

[original utterance of Ken:] Ii am cool.

b.

Ken said: “Ii am cool.”

c.

Ken said that hei is cool.

d.

Ken said that Ij am cool.

Not surprisingly, there have been doubts with respect to the assumption that direct and indirect quotations aim at representing an original utterance. First of all, the rationale of indirect quotation seems to be a licence for sloppy or inexact renderings of what has been said in the original utterance. There is some point in this general assumption, so let us put this issue aside and go into the (seemingly clear) case of direct quotation. Even here, the possibility of a verbatim reproduction of an original utterance is denied. For instance, von Roncador (1998) coins the term “nonliteral direct quotation” in this connection. The crucial cases are those cases in which possible or future attitudes or utterances are represented, as in (7a). Moreover, at least in German, there exists a class of elements like der und der, da und da, and so und so that have the property to substitute for referential expressions of the original utterance, as first observed by Harweg (1972); the point is that these dummy expressions (that are restricted to quotations) render an original utterance, but not in a literal manner. (7)

a.

Ken würde bestimmt sagen: „Das Projekt ist schwer zu verstehen.“ ‘Ken would certainly say: “The project is hard to understand.”’

b.

Ken würde bestimmt sagen, dass das Projekt schwer zu verstehen ist. ‘Ken would certainly say that the project is hard to understand.’

c.

Ken sagte: „Das und das ist schwer zu verstehen.“ ‘Ken said: “This and this is hard to understand.”’ [this and this referring to das Projekt]

Lying, direct/indirect quotation, and the verbatim assumption   

d.

   187

Ken sagte, dass das und das schwer zu verstehen ist. ‘Ken said that this and this is hard to understand.’

Does a quotation always require an original utterance, as we have postulated in our definition (3)? Possibly one should adopt a distinction between the rendering of merely possible attitudes or thoughts, and quotations in the strong sense. This can be left open since we are interested in lying while quoting, and since it is not possible to check the truth of a possible thought; we have no control over the factuality of the quoted material, e.g., whether Ken would say a certain thing. Harweg’s observation also calls for further explanantion. Since the original utterance is not rendered verbatim, there must be a reason for this. I see the reason in the triggering of conversational implicatures, for instance, that the speaker does not care for the precise description of the reported event, or that they do not recall what the original speaker said, etc., thus exploiting the maxims of Quantity and Manner. What is interesting for the analysis of lying is that the speaker has the license to use these expressions; however, if they know for sure that Ken used the words the project, but prefer to substitute this and this, then they may be said to act deceptively. The most influential attack on the assumption of verbatim reproduction (AVR) is found in Clark and Gerrig (1990) where they defend the general thesis that direct quotations are demonstrations. Demonstrations are illustrations or depictions of something by giving an example. They are distinguished from indicating and describing, two further modes of linguistic representation. According to Clark and Gerrig (1990), demonstrations have two important properties: First, they are ‘non-serious actions’ (Clark and Gerrig 1990: 766f., 770ff.), which means that the demonstrated action is not ascribed to the demonstrating person. Second, they are ‘selective depictions’ (Clark and Gerrig 1990: 767f., 774ff.), i.e., it is possible to select from the totality of information contained in an original utterance and focus on certain aspects of it. The second property motivates the rejection of the AVR. Clark and Gerrig (1990: 795–800) present several arguments in favor of their position. First, even the basic notion of verbatim reproduction is unclear. Second, because of memory limitations, no speaker is able to perfectly memorize an original utterance. Third, typical conversational repairs of speakers aiming at verbatim reproduction show their uncertainty in this respect. Fourth, the assumption of verbatim reproduction excludes translations, but it is “sheer philosophical imperialism” (Clark and Gerrig 1990: 799) to exclude them solely for this reason. Fifth, AVR denies the typical selectivity of reproductions, for instance, with respect to translations, change of dialects, sociolects and register, idiolects, information structure, style, citation of noise, impossible demonstrations (e.g., a speaking finger), free indi-

188   

   Lying and quotation

rect quotations, quoted thoughts, etc. Sixth, we even find integration of nonverbal elements, e.g., rude gestures or noises. These arguments are related to spoken language. Indeed, from the point of view of Clark and Gerrig (1990: 800), AVR is nothing but a prejudice that stems from the academic tradition with its focus on written language. Ordinary speakers, however, are in no way committed to these ideals. While these considerations show that the AVR may be taken cum grano salis, the assumption that one has access to the original utterance still grounds every theory of quotation. Even in everyday language, there is a good deal of reflection on the question of what someone said in reality, and to what extent the original utterance is misrepresented. Moreover, the citation practices in the academic world, in the political sphere, in jurisprudence, and in the media show that the verbatim assumption or the need to reconstruct the original utterance is important. Johnson and Lepore (2011) rebut the demonstration theory, showing that the arguments against AVR miss the mark. They adopt a version of a so-called ‘wording theory’; “a paradigm case of a strong wording theory is one that asserts the equivalence in [8]” (Johnson and Lepore 2011: 241): (8)

Strong wording theory For any direct quotation Q, ‘John says Q’ is true if and only if Q is a verbatim reproduction of an utterance made by the referent of ‘John’.

A weak wording theory, in contrast, is one that settles “on a direct quotation being true just in case it approximates the original speaker’s words (where what counts as an acceptable approximation varies with the context” (Johnson and Lepore 2011: 232). The examples in (9)–(13), presented as problematic cases even for weak wording theories by Wade and Clark (1993: 806–807), are shown to be no real problems (or, being problems for the demonstration theory, too). The only real problem is, according to Johnson and Lepore (2011), the case of translation, as in (12): (9)

You can’t say, ‘Well Daddy I didn’t hear you’. [non-existent utterances]

(10)

And I thought ‘well a I’m not going to’ you know ‘produce any sort of functional gaffes’. [unspoken thoughts]

(11)

Many people have come up to me and said ‘Ed, why don’t you run for the Senate?’. [generic utterances no particular person produced]

Lying, direct/indirect quotation, and the verbatim assumption   

   189

(12)

‘Well met, captain’ he said, quietly, in German. [translations]

(13)

Senatorial hair, flaring over the eras in authoritative gray waves, says ‘put me on Nightline’. [objects]

I side with Johnson and Lepore (2011) in arguing that these apparent counterexamples against wording theories are not really convincing. Examples (9)–(11) fall into the category of the quotation of possible thoughts or attitudes. (Hence, this category shares some properties with the problem of fiction discussed in section  5.3.) We have already put aside these cases, arguing that they should be excluded from the analysis of lying because no one asserted the “original thoughts”, so to speak. Johnson and Lepore (2011: 244) admit that example (12) poses a problem for strong as well as weak wording theories. Their view is that (12) “is true if, and only if, the referent of ‘he’ said the quotation ‘Well met, captain.’” Moreover, they “further claim you can say ‘Well met, captain’ in German, and you do so by … well, saying it in German.” In the first scenario, the speaker uttered something like Gut getroffen, Kapitän, so it really is conversationally implicated (by way of exploitation of the maxim of Manner) that a translation into English has taken place. In the second case, it is implicated that the English string Well met, captain! was used as part of a German discourse. It follows that the real problem is whether “saying” includes translations, so that speakers who say bonjour, hello, and guten Tag say the same thing. Example (13) is not problematic because senatorial hair is a metonymy. If this expression refers to a senator as a speaker, there is no problem for a wording theory to be involved. Johnson and Lepore (2011: 244) summarize their approach as follows: “Our view, then, is like a wording theory except we’ve replaced ‘reproduced verbatim’/’approximated’ in the truth conditions of direct reports with ‘said’. That is, for any quote Q, ‘Mary said Q’ is true if and only if Mary said Q; i.e. if Q is something Mary said; i.e. what Mary said was Q.” Moreover, they add: “We doubt we can provide a fully explicit analysis of ‘say’ any more forthcoming that we could one for ‘good’. But there’s no reason to suppose saying involves depiction. It’s a lot closer to verbatim reproduction, except for permitting cleaning-up and translation. It’s saying.” An important point is that a wording theory does not require that everything someone has said (including false pronunciations, mistakes, etc.) is quoted. A quotation is not a replication of an original event. Note that a normal literal sentence like There is a cat on the mat also presupposes a certain context, and that the speaker has to select certain aspects of the total event. Moreover, there is common

190   

   Lying and quotation

knowledge that an accurate quotation is, in principle, possible. “Speakers invariably defer when confronted with, say, recorded documents. They might resist conceding vis-à-vis corrected infelicities (unless they matter for the discourse at hand) but they will defer (concede error) for non-verbatim reports when the original is presented.” (Johnson and Lepore 2011: 245) In a similar vein, I argued in my (2007a) that it belongs to the quotation competence of ordinary speakers that they know the difference between correct quotation (one that respects the original utterance) and incorrect quotation (one that does not take the original utterance seriously, for whatever reasons). It is a social fact that several institutions such as the academia and the law have a professional interest in original utterances, as is true for everyday communication (for instance, what exactly has someone said about me?). Thus, we may safely conclude that lying by using an untruthful quotation is possible (and usual). Matters of inaccuracy further lying, as will be discussed in more detail in chapter 7. Let us now turn to the question which possibilities of lying are connected with both types of quotation, i.e., direct and indirect quotation. In (2) above, we already showed how direct quotations may be used for lying. In (14), we look at cases of indirect quotations. Here, we focus on the proposition of the matrix sentence and the embedded sentence, and we include the expression of attitudes towards these propositions. Note that misrepresentations of the original utterance can happen on a number of levels, e.g., phonetic, gestical, mimical, or motorical, but these cases are not covered by our definition in (3), and hence are neglected. (I also neglect cases in which a speaker asserts that OS said p, although OS said that nothing.) (14)

Original utterance by Ken: The project is hard to understand. a. Ken said that the project is easy to understand. b.

Ben said that the project is hard to understand.

c.

Ken stumbled pessimistically/shouted that the project is hard to understand.

d.

Ken said that the damn project is hard to understand.

e.

Alas, Ken said that the project is hard to understand.

The utterances in (14a) and (14b) are clear cases of lying according to our definition. Utterance (14c) is more problematic, since it demonstrates that in the matrix sentence, there are many more verbs and verbal constructions possible beside the neutral verb of saying to say. The speaker of the quotation has a certain latitude for interpreting the illocutionary act performed by the original speaker. For instance, if it

Lying, direct/indirect quotation, and the verbatim assumption   

   191

is clear in the context of the original utterance that Ken stated that the project is hard to understand, while the speaker of the quotation asserts that Ken issued a warning that the project is hard to understand, then the speaker lied. By the same token, the characterizations given in (14c), i.e., stumbled pessimistically and shouted, may be untruthful characterizations of what Ken really did when performing the original utterance. This poses no problem for our definition of lying because the description of the respective illocutionary act is part of the proposition of the matrix sentence. Example (14d) contains the expressive adjective damn in the embedded sentence. Since this expressive device was not contained in the original utterance, (14d) gives the false impression (at least according to wording theories) that Ken has a hostile attitude toward the project and, moreover, that he did use that particular word at all. Similarly, the particle alas in (14e) conveys an expressive meaning, but this time, since the expressive attitude can only be ascribed to the speaker, there is no lie involved because the speaker is free to express their attitude to what Ken is reported to have said. Intuitively, cases like (14d) are lies. If these cases are to be analyzed as lies, we have to extend our definitions so that they comprise attitudes towards p. Alternatively, we have to define p in a way that p comprises propositional as well as non-propositional yet literal meanings. There are several analyses to be found in the literature, for example, Potts (2005, 2007) who characterizes damn as an expressive device connected with a conventional implicature, or Predelli (2010) who discusses the interjection alas. If an approach like Potts’ is accepted, one may think about untruthful conventional implicatures in analogy to our analysis of untruthful conversational implicatures. Let us now turn to the question whether conversational implicatures may be quoted and are hence subject to lying. Recall my plea for including intentionally false implicatures in a comprehensive notion of lying, as developed in section 4.5. This is in line with Bach (2006: 27) who states that an implicature is part of the truth-conditional content of an utterance. (See his Number 7 of the “top ten misconceptions”, stating that “implicatures are not part of the truth-conditional contents of utterances”.) To begin with, it can be shown that quotations are transparent for conversational implicatures. Let us consider the classical case of a particularized conversational implicature presented in Grice (1989a: 24): (15)

[Context: A and B are talking about a mutual friend, C, who is now working in a bank] A: How is C getting on in his new job? B: Oh quite well I think; he likes his colleagues, and he hasn’t been to prison yet.

192   

   Lying and quotation

Grice mentions possible conversational implicatures like (i) “C is the sort of person likely to yield to the temptation provided by his occupation”, and (ii) “C’s colleagues are really very unpleasant and treacherous people”. For the sake of simplicity, let us assume that the intended and derived implicature is ‘C ist potentially dishonest’. Suppose that later, A reports this piece of discourse to D by using an indirect quotation. Then it is possible to convey the implicature intended by B in the following way: (16)

B said that C is quite well getting on in his new job and that he likes his colleagues and has not been to prison yet. +> ‘C is potentially dishonest.’

The quotation is sincere, no lying is involved. It is similar in the case of the Mate and the Captain if someone reports the original utterance (logbook entry) as in (17): (17)

The mate wrote into the log that the captain was sober. +> ‘The captain was mostly drunk.’

Considering the possibility to convey implicatures by using quotations, we are faced with two questions. First, is the speaker of the quotation responsible for the transmission of the conversational implicature, either implicitly or explicitly, or do they simply count on the interpretive capacities of the hearer, thus letting the derivation of the respective interpretative hypothesis (for instance, the writer of (17) wanted me to derive that the mate wanted to implicate that the captain was mostly sober) be completely up to the hearer or reader? Second, is the conversational implicature solely attributed to the speaker of the original utterance, or is it possible to attribute it (at least to a certain degree) also to the speaker of the quotation who might be in a position of agreeing to the implicature or distancing themself from it? Thus, a hearer of the quotations in (16) and (17) might ask the speaker So is q your opinion, too?, with q standing for the conversational implicature. Similarly, in (14d), there is also an interpretation available in which the derogatory attitude connected with damn is ascribed to the speaker. If the hearer knows that the speaker likes the project, the respective reading is kicked out by this contextual knowledge. But if the hearer does not have this particular knowledge, the speaker should distance themself from the conveyed negative attitude, for instance, by adding This is his opinion, not mine, or a similar cancelling phrase. The derivation of an implicature is of course dependent on some knowledge of the situation of utterance and background knowledge related to the situation. To quote an original utterance without attributing to the hearer any knowledge

Lying, direct/indirect quotation, and the verbatim assumption   

   193

about the context of the original utterance does not make any sense. Therefore, I will assume in the following analysis that a conversational implicature belongs to the original utterance, whether it remains implicit or is made explicit, and this opens up a certain lying potential for the speaker of the quotation. We distinguish three cases: (18)

a.

The conversational implicature q of the original utterance is suppressed in the context of quotation. Instead, the conversational implicature -q is conveyed.

b.

The conversational implicature -q of the original utterance is suppressed in the context of quotation. Instead, the conversational implicature q is conveyed.

c.

The original utterance contains the conversational implicature q, but the speaker of the quotation implicates (or otherwise manages to lead the hearer into thinking) that it was the conversational implicature r.

Let us illustrate these cases, taking the quantity implicature (GCI) based on the scale . (19)

r +> q; “p (r)” +> -q a. Original utterance (by Ken): Some projects are good. +> ‘Not all projects are good.’ b.

Quotation (by Barbie, said to Ben): Ken said that some projects are good but not all of them +> ‘in fact, all’

Imagine a scenario in which the speaker of (19b), say Barbie, opts for as much projects as possible. The addressee of (19b), Ben, however, is known to be a principal opponent of these projects. The speaker of the original utterance, Ken, is a consultant. Ken intended that Barbie derived that not all projects are good. This, however, is not what is in the interest of Barbie, and so she wants to create the impression on Ben that Ken was cautious about judging and that he possibly judges all the projects as good. Hence, Barbie suppresses the conversational implicature of Ken’s original utterance; in fact she conveys the opposite of it. This is, according to my approach to lying that includes untruthful implicatures, a lie. The speaker of (19b) knows that the implicature of the original utterance (the original implicature) was q, but she transmits -q. This procedure works also works the other way around.

194   

(20)

   Lying and quotation

r +> -q; “p(r)”+> q a. Original utterance (by Ken): Some projects are good. +> ‘in fact, all’ b.

Quotation (by Barbie): Ken said that some projects are good. +> ‘Not all projects are good.’

In this case, Ken’s original utterance is an understatement; however, Barbie chooses to ignore this and acts in a way that leaves the impression on Ben that the standard implicature must be applied. Again, Barbie is lying. This pattern of lying works because there is a scale mutually known to the participants. In most cases, the set of possible implicatures is not fixed and varies with context assumptions (cf. Grice 1989a: 39). Take the case of the definite article. If some one utters Pete is meeting a woman this evening, then it is implicated (via the maxim of Quantity) that the referent of a woman was not Pete’s wife, mother, sister, or another close relative (Grice 1989a: 37). Similarly, in the original utterance (21a) the implicature is triggered that the Porsche was not owned by Barbie. (21)

r +> q; “p(r)” +> s a. Original utterance (by Ken): Barbie was driving a Porsche. +> ‘Barbie was not the owner.’ b.

Quotation: Ken said that Barbie had been driving a Porsche. +> ‘Barbie is rich.’

If the speaker reports the original utterance in a way that the implicature ‘Barbie is rich’ is derived (and this may include the assumption that she is the owner of the Porsche), then the speaker manipulates the complete content of the original utterance. In (22), I propose a definition for untruthful quotation that covers the aspect of those implicatures intended by the original speaker and manifested in the context of the original utterance. I proceed from the assumption that it is a sincerity condition for quotations that the speaker is responsible for correctly reporting the complete content of the original utterance. This must not happen explicitly, since the reporter is licensed to assume that the addressee has certain contextual knowledge with respect the original situation as well. In particular, the reporter must be careful with respect to the implicatures their report may convey. Manipulating these implicatures is a prime strategy of lying with quotations.

Lying and mixed quotation   

(22)

   195

Lying by untruthful quotation S lied at time t by with the quotation p(r), iff (a) they asserted p(r) but actively believed that not p(r), (b) or they misrepresented conversational implicatures of p(r).

‘Misrepresentation’ is a practical short-hand for the cases covered in (19)–(21). Note that it is not interesting whether the original utterance was a lie. If the speaker of the quotation wants to trigger untruthful implicatures not related to the original utterance, our normal analysis for untruthful implicatures applies. For instance, the speaker utters Ken said to some people … when they know for sure that Ken addressed all of them. (With respect to the cases discussed in (19– 21), the restriction to r would suffice.)

6.3 Lying and mixed quotation With mixed quotation, only a part of the original utterance is directly quoted. Thus, all utterances in (23) are mixed quotations: (23)

Original utterance by Ken: The project is hard to understand. a. Ken said that “the” project is hard to understand. b.

Ken said that the “project” is hard to understand.

c.

Ken said that “the project” is hard to understand.

d.

Ken said that the project “is hard to understand”.

e.

Ken said that the project “is” hard to understand.

f.

Ken said that the project is “hard” to understand.

g.

Ken said that the project is hard “to understand”.

The material in quotation marks is quoted from the original utterance. Whether the material outside the quotation marks is the same as in the original utterance or whether it only resembles it to a certain degree may vary. The use of quotation marks in (23) is similar to focussing. In the original situation, there were alternative expressions that had not been chosen by the original speaker. The speaker of the quotation thus draws attention to the particular choice of the original speaker, thus disregarding their alternative phrasings. For instance, such an alternative could be this in (23a), plan in (23b), might be in (23d), or to develop in

196   

   Lying and quotation

(23g). The focus on “is” in (23e) is most likely a verum focus (cf. Höhle 1992 and the discussion in section 7.3). Pragmatic approaches to mixed quotations like Recanati (2001), Saka (1998, 2003), and Stainton (1999) assume that the pair in (24) has the same meaning, i.e., both sentences have the same truth-conditions. (24)

a.

Barbie said that the song is hard to sing.

b.

Barbie said that the song is “hard to sing”.

That the second utterance is usually understood to mean that Barbie in fact uttered the words “hard to sing” is a pragmatic matter; it is nothing that is directly expressed by this utterance. Semantic approaches, in contrast, endorse the view that this difference is of a semantic nature, i.e., it is part of the truth conditions for (24b) that Barbie used these particular words. Reimer (2003: 178ff.) discusses some pro and con arguments of the pragmatic approach. Look at the examples in (25) where (a)–(d) are additions to the utterance Quine said that quotation “has a certain anomalous feature”. (25)

Quine said that quotation “has a certain anomalous feature.” a. … # However, those were not the words he used. b.

… However, he didn’t use those words; I think he said it in German.

c.

… However, I doubt Quine ever used a pretentious word like “anomalous.” He was a concise writer and so probably said that quotation has an “odd” feature.

d.

… – as you are so fond of putting it.

While example (25a) is intended to show that the assumption that the original speaker used are exactly the words enclosed in the quotation marks, thus supporting the semantic approach, the other examples (25b–d) are attempts at suspending this assumption. However, these additions are normal variants of cancellability tests and in the end, they amount to clarifications. Reimer (2003: 181) presents an own example intended to show evidence in favor of the pragmatic approach: (26)

a.

She said that life is

difficult to understand.

b.

She said that life is

difficult to understand.

Lying and mixed quotation   

   197

Big font indicates loud voice, small font indicates soft voice. In this case, it is clear that respective properties are not cancellable, yet they do not contribute compositionally to the truth-conditions of these sentences. In a similar vein, quotation marks could be understood as devices that contribute to sentence meaning, albeit not in a truth-conditional manner (Reimer 2003: 182). This would be in line with Recanati’s (2001) approach. As far as I see, the analogy is not really convincing because, by definition, loudness of voice is not a property of a proposition. But we can take it for granted that there is non-propositional meaning, for example, the meaning of interjections or modal particles. What is missing is the relation to indirect quotations. Thus, the cancellability test also fails when applied to indirect quotations: (27)

Quine said that quotation has a certain anomalous feature. …# However, those were not the words he used.

(Note that the addition in (27) is only acceptable if understood as a correction of the assumption that Quine had used exactly these words. This is similar in the nice example provided by Claudia Poschmann (p.c.): In fact, Quine says that quotations are quite strange sorts of beasts. Of course, he never used these words. As a concise writer, he said: “Quotation has a certain anomalous feature”.) This shows that the verbatim assumption plays a role in indirect quotations, too. Hence, misrepresentation of the original utterance is possible with both types of quotation. In general, quotation marks have the function to highlight a certain string as a part of the original utterance; with direct quotations, it is the whole utterance, with mixed quotations, it is a part (a fragment) of the original utterance. Indirect quotations, as I have argued, are related to an original utterance, too, but here we do not get any clues to exactly determine which parts of the original utterance are quoted. Nevertheless, it must be possible to distinguish between an indirect quotation that has no relation whatsoever to the original utterance, and fair and transparent renderings of the original utterance (see section  6.5). Highlighting strings of the original utterance may have further effects beyond indicating that something has been said by an original speaker, for example, in the case of scare quotes (see section 6.4). I conclude, then, that quotation marks may be regarded as pragmatic indicators, as long as it is clear that in the case of direct and indirect quotation, their task is to mark the quoted material as being the exact wording used by the original speaker (cf. Recanati 2001, Gutzmann, and Stei 2011).The crucial point is what motivates the speaker to select one of the variants in (23a–g). Recall Levinson’s (2000: 76) version of the maxim of Quantity:

198   

(28)

   Lying and quotation

Q-Principle Speaker’s maxim: Do not provide a statement that is informationally weaker than your knowledge of the world allows, unless providing an informationally stronger statement would contravene the I-principle. Specifically, select the informationally strongest paradigmatic alternate that is consistent with the facts. Recipient’s corollary: Take it that the speaker made the strongest statement consistent with what he knows […].

Applied to mixed quotation, we may ask why the speaker selected the mode of mixed quotation if they could have selected a (complete) direct or indirect quotation. We assume that the speaker has certain reasons for suppressing or highlighting certain information. This may have several functions (see Clark and Gerrig 1990: 792–793), amongst them (a) focussing on a certain expression because the way of putting something is important (see Grice’s maxim of Manner), (b) stylistic improvement because representation of the original utterance would lead to incomprehensibility or unintelligibility, (c) protection of the original speaker, for instance, if the original speaker used embarrassing words that are left out form the quotation, (d) defamation of the original speaker, for instance, if the focussed material is of outstanding embarrassment. Which of these functions are instantiated in a concrete piece of discourse has to be pragmatically inferred. The respective implicatures are implicatures triggered by the quotation; hence, the speaker of the quotation is responsible for these implicatures, not the speaker of the original utterance. However, the speaker of a mixed quotation may draw the hearer’s attention to certain implicatures of the original utterance as, for instance, in Ken said that it is a “not unimportant” project +> ‘a rather important project’. Is it possible to lie by using mixed quotation? The aforementioned speaker implicatures are not related to the truth or falsity of a propositional content, so they are not captured by our definition of untruthful implicature. It is more difficult to decide whether the utterances in (29) could count as lies: (29)

Original utterance by Ken: The project is hard to understand. a. Ken said that the enterprise is “hard to understand”. b.

Ken said that the project is “not easy to understand”.

On the one hand, it could be argued that (29a) is no lie because it is not asserted that Ken literally spoke of enterprise; moreover, since enterprise may be regarded as synonymous to project, there is no difference in meaning. On the other hand,

Lying and scare quotation   

   199

the speaker of the quotation is not released from the obligation to speak truthfully solely because of the fact that they used a mixed quotation; in indirect quotation, too, there is a presumption that the speaker represents the original utterance as correctly as possible. Seen from this angle, (29a) could be a lie, at least insofar as there might be a deception involved with respect to the use of enterprise instead of project. It goes without saying that in everyday talk we would not suspect that the speaker was lying because we grant a certain leeway against too strict requirements of verbatim quotation. If (29a) is counted as a lie, there is, in comparison to (29b), a difference such that (29a) is a “weak” lie, while (29b) is a “strong” lie. In the latter case, though hard to understand and not easy to understand may be regarded as synonymous, we clearly have a case of misrepresentation. And, mistakes excluded, this amounts to lying according to our definition.

6.4 Lying and scare quotation Scare quotation goes together with conveying a certain attitude of the quoting speaker towards the material enclosed in quotation marks. Thus, a quotation like (30a) can either be understood as a mixed quotation, as in (30b), or as a scare quotation, as in (30c). In the latter case, the speaker of the quotation wants to suggest that the project does not deserve the title of a “project” since it is, according to them, only a mini-project or a pseudo-project, etc. What the suggestion of the speaker exactly is has to be derived from the context of utterance. (30)

Original utterance by Ken: The project is hard to understand. a. Ken said that the “project” is hard to understand. b.

Mixed quotation: Certain reasons led the speaker to highlighting project.

c.

Scare quotation: Speaker thinks that the project does not deserve this name.

The fact that there is pragmatic ambiguity in (30a) suggests that there is a connection between both interpretations, i.e., mixed quotation and scare quotation. Note that we may have scare “quotations” without any original utterance; these cases are excluded by our definition of quotations since it is clear that not everything included in quotation marks is related to an original utterance. In the scare quote interpretation of (30a), the particular attitude towards the project is attributed to the speaker of the quotation. In direct quotation, the scare

200   

   Lying and quotation

quote attitude is attributed to the speaker of the original utterance. Therefore, in (31b), we have a case of lying, but not in (31a): (31)

Original utterance by Ken: The “project” is hard to understand. a. Ken said: “The ‘project’ is hard to understand.” b.

Ken said: “The project is hard to understand.”

Furthermore, (32a) is a lie because a scare quotation is reported that was lacking in the original utterance. (32)

Original utterance by Ken: The project is hard to understand. a. Ken said: “The ‘project’ is hard to understand.”

Predelli (2003a, b) develops a theoretical approach, according to which scare quotation is analyzed as an interplay of so-called ‘message’ and ‘attachment’. He discusses the following examples (Predelli 2003a: 2–4; his examples are given with single quotes): (33)

in offset printing ‘printing proofs’ of illustrations come from the darkroom, not the proof press

(34)

had it not been for Bryce, the ‘coppers nark’, Collins would have made his escape

(35)

national greed has disguised itself in mandates to govern ‘inferior’ races

(36)

the ‘debate’ resulted in three cracked heads and two broken noses

(37)

this remarkable piece of ‘art’ consists of a large canvas covered with mud and old bus transfers

(38)

myths of ‘paradise lost’ are common in folklore

(39)

life is ‘the farce which everybody has to perform’

(40)

life is ‘what happens while you are making other plans’

In (33), the writer assumes or knows that the word in quotation marks is not really adequate. In (34), the writer uses a slang expression, i.e., the writer excuses

Lying and scare quotation   

   201

themself. In (35), the speaker wants do express a reserved attitude towards the content of the word in quotation marks. Sentences (36) and (37) display sarcasm. In (38), the quotation marks mark a phrase that is presupposed to be mutually known, and in (39) and (40) the quotation marks mark quotations whose author is presupposed to be known (Arthur Rimbaud, John Lennon). In all these cases, the expressions in quotation marks are used, not mentioned. The quotation marks signal that the quoted expressions have certain specific properties, be it that they are used in an unusual way (33–37), or that they must be attributed to a mutually known source (38–40). (Let me add that I would not classify (39–40) as scare quotes at all. Note that Klockow (1980), who discusses scare quotes at length, also did not include examples of this type.) According to Predelli (2003a: 7), the message is the content of the utterance, while the attachment is the specific meaning that is conveyed by the use of the quotation marks. Hence, the quotation marks in (33–40) function as ‘attachment triggers’. These are comparable with expressions like but and therefore, which, according to Bach (1999), are also connected with information that surmounts the message of an utterance. For Predelli, it is important that the attachment is a matter of the semantic profile of an utterance: (41)

Truth-functionality of attachments „[…] the content triggered by the quotes is truth-functionally relevant, in the sense that, at least with respect to some contexts, sentences such as [(33)–(40)] display a truth-functionally different profile than their counterparts without quotation marks.“ (Predelli 2003a: 18)

The reason for this claim is that negations may relate to attachments. For instance, if an American speaker writes We don’t watch ‘colour TV’ here (thereby insisting on the difference between English and American spelling), it is not the message that is negated, but the attachment. (Note that this is an example for metalinguistic negation that does not really prove that there is need for the notion of an attachment.) The major drawback of such an analysis, as I see it, is of course the introduction of somewhat idiosyncratic terminology. The notion of ‘attachment’ appears to be similar to the notion of conventional implicature (Horn 2008: 24), but it is not plausible that the attachment meanings in (33)–(40) are conventional implicatures. Recently, Potts (2005, 2007) has argued for the significance of conventional implicature. Drawing on an exegesis of Grice’s work, his core criteria are (a) that “CIs are part of the conventional meanings of words”, (b) that “CIs are commitments, and thus give rise to entailments”, (c) that these commitments “are made by the speaker of the utterance ‘by virtue of the meaning of’ the words he

202   

   Lying and quotation

chooses”, and (d), that they “are logically and compositionally independent of what is ‘said (in the favored sense)’ […]” (Potts 2005: 11). But scare quote meaning is never part of the conventional meanings of words, and for this reason I doubt that conventional implicature (let alone attachment) is the correct category label here. Furthermore, as Gutzmann and Stei (2011) point out, there is the problem that in ironical scare quotes the total semantic profile might contain contradictory information, e.g., in Peter’s bagels are “fresh”, the information that Peter’s bagels are fresh and not fresh at the same time. Hence, Predelli has to account for the substitutivity of ironies. And finally, it has to be spelled out how the particular interpretations displayed in Predelli’s examples may be connected with attachments if understood as conventional implicatures. Instead of treating scare quotes as attachments or conventional implicatures, their treatment as conversational implicatures is far more promising. This idea has already been proposed by Klockow (1980) and is elaborated in Gutzmann and Stei (2011). (Note that Klockow finally abandoned this idea because it seemed too narrow to him to catch all the diverse effects connected with scare quotation.) If we have an original utterance, i.e., in the case of an overlapping of mixed and scare quotation, we invoke the maxim of Quantity (see 28) because using quotation marks in such cases is intuitively “informationally stronger” than not using them. If we do not have an original utterance, an implicature may be derived while making use of the maxim of Quantity (Q-principle), the maxim of Relevance (why are quotation marks relevant?) or the maxim of Manner (M-principle) (an expression in quotation marks is “more marked” than the same expression without quotation marks). So lying by using scare quotation is possible, for instance, in cases of irony and sarcasm as in (36) and (37); recall our discussion of lying while untruthfully implicating. To conclude this section, I will shortly go into free (in)direct quotation, emphatic quotation, and pure quotation. In the case of free (in)direct quotation, a thought is attributed to a person. In order to find out whether free (in) direct quotation may count as a lie, we should have access to the truth or falsity of this thought, but we have not. Our definition of untruthful quotation requires an original utterance, and this we do not have in the case of (in)direct quotation. It is similar with emphatic quotations (cf. Abbott 2003) and pure quotations: here, we do not rely on original utterances. Admittedly, we may be deceived when “project” is written on the door leading to the toilet, but this deception is possible without any quotation marks. Pure quotations are not related to original utterances; of course, “Monosyllabic” is monosyllabic may be a lie, but this is not a lie according to our definition in (22) because no original utterance has been quoted.

Fairness and transparency   

   203

6.5 Fairness and transparency While I have insisted that the relation to an original utterance and the verbatim assumption are important for the understanding of a quotation, I side with Johnson and Lepore (2011: 245) in permitting “some degree of fixing up”. In order to pinpoint the parameters speakers observe in deciding how much “fixing up” is tolerable, we probably need further pragmatic principles. Thus, Lang (1982:  318) draws a distinction between two properties of quotations, namely fairness and transparency. The principle of fairness is defined as follows (my translation, JM): (42)

Fairness (Lang 1982: 318) A quotation is fair (a) if it allows to reconstruct the propositional content and the attitudinal frame of the original utterance, and (b) if it phrases the reference to the communicative sense in such a way that in the quotation, the intention of the original speaker may be distinguished from its interpretation or evaluation by the reporting speaker (the speaker of the quotation).

Lang characterizes “fairness” as a semantic/pragmatic property of the quotation, but one might as well understand “fairness” as a norm guiding the speaker of the quotation. Deliberately unfair quotation amounts to misrepresentation and lying. The second property mentioned by Lang is transparency: (43)

Transparency (Lang 1982: 318) A quotation is transparent if the linguistic structure of the quotation allows reconstructing the linguistic structure of the original utterance.

Transparency is conceived of as a syntactic property of the quotation. Since it is the speaker who is responsible for the selection of particular syntactic structures, this property may also be viewed as a pragmatic criterion. If the speaker of a quotation observes the norm that they should quote as verbatim as possible, then they also should look for transparent structures. Lang postulates that the two principles should be thought of as being independent from each other; applied to our standard example, this yields the following set of combinations:

204   

(44)

   Lying and quotation

Combinations of fairness and transparency Original utterance by Ken: Das Projekt ist schwer zu verstehen. a. Ken sagte: „Das Projekt ist schwer zu verstehen.” ‘Ken said: “The project is hard to understand.”’ [+fair, +transparent] b.

Ken sagte, dass das Projekt schwer zu verstehen sei. Ken said that the project hard to understand isSUBJ. ‘Ken said that the project was hard to understand.’ [+fair, –transparent]

c.

Ken moserte: „Das Projekt ist schwer zu verstehen.“ Ken grizzled: “The project is hard to understand.” [–fair, +transparent]

d.

Ken moserte, dass das Projekt schwer zu verstehen sei. Ken grizzled that the project hard to understand isSUBJ. ‘Ken grizzled that the project was hard to understand.’ [–fair, –transparent]

We assume that sagen (to say) is a fair verb of saying since it is, in a sense, neutral and prototypical, cf. (42b). In addition, we may assume that direct quotation is maximally transparent. Thus, (44a) is marked [+fair, +transparent]. In contrast, while (44b) is fair, there is a problem with respect to transparency. Note that in German, indirect quotation may be marked by the subjunctive. In everyday language, mostly the indicative is used; however, in the language of newspapers, for instance, indirect quotation is marked by the subjunctive. Now imagine a case in which the original utterance already uses the subjunctive. Then we face the problem that we do not know whether the subjunctive is merely used to indicate quotation, or whether the subjunctive was already contained in the original utterance, hence making the quotation verbatim. To this extent, the quotation is nontransparent, yet fair. In contrast, (44c) is unfair because of the verb of saying mosern having a pejorative meaning. If the speaker of the quotation thinks that the utterance of Ken is not correctly described by using the verb mosern then they are lying. Hence, (44c) is unfair, yet transparent. Finally, (44d) is unfair as well as not transparent. The criteria of fairness and transparency are certainly a step in the right direction, when it comes to analyzing correct and incorrect quotation. Let us finally have a look at authentic examples that include translations from English into German; the English original has already been discussed as an example of bullshitting in section 5.6. On January 23th 2012, the German model and business-

Fairness and transparency   

   205

woman Heidi Klum and her husband, the singer Seal, declared their separation. In (45)–(47), there are three reports about his fact taken from different sources: (45)

In a joint statement to E! News, the couple said, “While we have enjoyed seven very loving, loyal and happy years of marriage, after much soulsearching we have decided to separate. We have had the deepest respect for one another throughout our relationship and continue to love each other very much, but we have grown apart. This is an amicable process and protecting the well-being of our children remains our top priority, especially during this time of transition. We thank our family, friends, and fans for their kind words of support. And for our children’s sake, we appreciate you respecting our privacy.” [http://prweb.com/releases/actingauditions/modelingauditions/ prweb9142093.htm] [01/27/2012]

(46)

„Obwohl wir sieben sehr liebevolle, loyale und glückliche Ehejahre genießen durften, haben wir uns nach reiflicher Überlegung dazu entschlossen, uns zu trennen“, hieß es in einer Erklärung gegenüber dem US-Magazin „People“. „Wir hatten den aufrichtigsten Respekt füreinander während unserer Beziehung und lieben uns nach wie vor sehr, doch wir haben uns auseinandergelebt.“ [http://www.freiepresse.de/NACHRICHTEN/PANORAMA/Heidi-Klumund-Seal-trennen.sich-artikeln78884552.php] [01/27/2012]

(47)

„Wir haben sieben liebevolle, treue und glückliche Ehejahre verlebt, aber nachdem wir in uns gegangen sind, haben wir die Trennung beschlossen“, lautete die Erklärung, die das Paar am Montag (deutscher Zeit) an einige Medien gab. Man habe sich „auseinandergelebt“. [Wiesbadener Kurier, 01/24/2012, p. 15]

While there are obvious differences in the translations – for instance, Engl. loyal has been translated with Germ. loyal in (46) and treu in (47), or after much soulsearching has been translated with nach reiflicher Überlegung in (46) and in sich gehen in (47) – we would still consider these translations as fair. However, they are not transparent because it is not possible to reconstruct the original utterance. To grow apart is translated with sich auseinanderleben in both (46) and (47); however, in (46) it is part of a whole quotation, while in (47) it is mixedquoted, with the possibility of interpreting it as a scare quote. (The latter interpretation would make sense, considering the bullshitting quality of this press

206   

   Lying and quotation

release, where the normal reader would ask themself while reading it why the couple separates at all, when it is true  – as they affirm  – that they love each other.) Even when no translation is involved, it can be shown that one and the same original utterance is quoted in different ways. However fairness and transparency are instantiated in authentic cases, it has to be admitted that there is no simple method of detecting these properties, and that these properties appear to be of a scalar nature. Scalarity also appears to play a role with respect to the degree of commitment versus distance to the quoted material. In German, the most prominent device for indicating a quotation is the subjunctive; however, this device is also seen as a means to express distance to the quoted material. Furthermore, there are many different lexical devices, for instance, verbs of saying, evaluative adverbs, etc. that indicate degrees of commitment or distance. In this section, we did not go into analyzing the whole spectrum of these devices. Instead, we focussed on relatively clear criteria for analyzing a quotation as a lie. In the next section, we will shortly deal with a more complicated case, the German modal verb wollen used in a quotative construction.

6.6 The case of the German quotative wollen Occasionally, one can read statements like the ones in (48) and (49) in German newspaper reports: (48)

Auch in Sex-Clubs in der US-Stadt Atlanta 1996 und im slowakischen Bratislava 2008 will der König nie gewesen sein. Also in sex clubs in the US-city Atlanta 1996 and in Slovakian Bratislava 2008 wants the king never beenPART.PERF. beINF. ‘The king also denied to have been in sex clubs in the US-city in Atlanta in 1996 and in Slovakian Bratislava in 2008.’ [Wiesbadener Kurier, 05/31/2011, p. 17]

The case of the German quotative wollen   

(49)

   207

Weder die Zwölfjährige, noch deren Eltern oder andere Menschen in ihrer Umgebung wollen von der Schwangerschaft etwas bemerkt haben, sagte ein Sprecher des Hilfsdienstes. Neither the twelve-year-old nor her parents or other people in her surrounding want of the pregnancy something noticedPART.PERF. haveINF., said a speaker of the emergency service. ‘The twelve-year-old as well as her parents and other people in her surrounding denied to have noticed signs of the pregnancy, a speaker of the emergency service said.’ [Wiesbadener Kurier, 03/30/2011, p. 17]

Both examples display a construction with the modal verb wollen plus Part. Perf. plus infinitive of the auxiliary sein or haben. Curme (1922: 322) already characterizes this use of wollen as being connected with “a claim that someone makes”, and he translated the example Der Zeuge will den Angeklagten gesehen haben with ‘The witness claims to have seen the defendant’. Since examples like (48) and (49) are always understood as reports of an original speaker, Letnes (2002: 101) proposes to subsume cases like this under the label of a reporting (‘referierend’) use of wollen, as distinct to the mere inferential (epistemic, subjective) use you find in sentences like Der Zeuge kann/muss/dürfte den Angeklagten gesehen haben (…can, must, might have seen …). I will endorse this terminological proposal, but switch to ‘quotative’ instead of ‘reporting’. In addition to quotative wollen, Letnes (2002) refers to ‘pretending’ (‘verstellend’) wollen, as in (50): (50)

Ich will es nicht gesehen haben. Curme (1922: 322): ‘I will claim, pretend that I did not see it.’

Letnes (2002: 101) defends the hypothesis that quoting and pretending wollen are related but nevertheless distinct categories of use. He assumes that both uses are based on the core meaning of wollen ‘will of the subject referent’; the readings as reporting vis-á-vis pretending are dependent on the context of use. The main difference between the variants is that the speaker of the pretending wollen without any doubt considers p as false. For instance, it is clear in the context of (50) that the speaker has indeed seen a certain event. While agreeing with these assumptions, I will try to show that the respective meanings are best understood as conversational implicatures; it is not so much a “sceptical reading”, as Mortelmans and Vanderbiesen (2011: 68) seem to assume. Moreover, I will show how untruthful implicatures may arise in the case of quoting wollen, thus embedding this case in the overall study of lying

208   

   Lying and quotation

and quotation. (For the case of reportative sollen, see the recent analysis by Faller 2012.) Let us have a closer look at (50). Imagine a situation in which a teacher watches pupils throwing away their chewing gums. In this situation, it is evident for the participants that the teacher was a witness of this forbidden behavior. Hence, it is implausible that the speaker pretends to have seen nothing. This is a blatant violation of the maxim of Quality. Hence, an implicature arises that may be paraphrased as ‘I wished that it would be the case that I did not see x.’ Thus, x is negatively evaluated. The speaker has seen x but it would have been better if they had not seen x. The utterance counts as a warning that x should not be repeated; it is a promise that x will not be reported to others. It is similar with other utterances cited by Letnes (2002) and occasionally mentioned in the literature, all of which require special situations in which retractions play a role: (51)

a.

Ich will nichts gesagt haben. I want nothing said have. ‘I didn’t say anything, o.k, (but…).’

b.

Das will ich auch gemeint haben. That want I also meant have. ‘I was going to say!’

c.

Das will ich Ihnen auch geraten haben. That want I youPOL. also advise have. ‘You’d damn well better!’

In (51a), the speaker retracts something they have just uttered, e.g., an insult; in (51a) and (51c), the addressee has retracted an utterance or corrected their misbehavior. Note that quotative wollen and pretending wollen behave differently with respect to the sentence adverb angeblich ‘allegedly’ (see Socka 2012 for an analysis of this lexeme): (52)

Auch in Sex-Clubs in der US-Stadt Atlanta 1996 und im slowakischen Bratislava 2008 will der König angeblich nie gewesen sein.

(53)

*Ich will es angeblich nicht gesehen haben.

The case of the German quotative wollen   

   209

This has to do with the third-person versus first-person perspective involved in the quotative versus pretending usages. Now consider the examples in (48) and (49). In these quotative uses of wollen, there arises a conversational implicature, too. Here, the speaker of the quotation often insinuates that the original speaker is untruthful, hence a liar. For instance, (49) is a part of a newspaper report dealing with the pregnancy of a twelve-yearold pupil who gave birth to a baby while being on a school trip. While it is not excluded that the young mother, her parents, and others were really surprised and were not aware of the pregnancy, this could have been judged as implausible by others. The respective implicature, arising by an apparent violation of the maxim of Quality, may be cancelled indeed: (54)

Weder die Zwölfjährige, noch deren Eltern oder andere Menschen in ihrer Umgebung wollen von der Schwangerschaft etwas bemerkt haben, sagte ein Sprecher des Hilfsdienstes. Dies ist auch vollkommen plausibel, wie Experten bestätigt haben. (This is totally plausible, as experts confirmed.)

In the context of (48), the article continued in (55): (55)

Zu den vielen angeblichen Zeugen meinte er: „Leider, leider, solche Clubs besuchen viele Menschen. Und die können nicht immer auseinanderhalten, wer da gewesen ist und wer nicht.” (Commenting on the many alleged witnesses he said: “Unfortunately, unfortunately, such clubs are visited by many people. And these are not always able to distinguish between those who were there and those who were not.”)

Thus, the implicature that the king did not say the truth could be suspended if the reader trusts the king’s comments. According to my analysis of lying while saying the truth, the implicature triggered by quotative wollen may also be untruthful. This may happen in the context of gossip. Suppose that someone has evidence that the twelve-year-old mother and the King of Sweden, Carl Gustaf XIV., were innocent with respect to the pregnancy or visits in sex clubs. Then, by quoting them by using the quotative wollen, it is possible to discredit them without any danger for the speaker of the quotation to be caught in the act of lying. After all, they did nothing else than reporting the defensive utterances of the accused persons. What is astonishing is that the language allows for such a possibility, proffering a complex verbal construction.

210   

   Lying and quotation

6.7 Conclusions In this chapter, I have argued that there are a number of procedures for lying with quotations. My main claim was that a quotation is a lie if the speaker of the quotation deliberately misrepresents the content of an original utterance or the attitude towards that content held by the original speaker. This also includes the untruthful representation of implicatures that were triggered by the original utterance. On the one hand, we need, contra the demonstration theorists, the verbatim assumption to make these deceptive procedures crystal clear. On the other hand, because the speaker of a quotation has a certain leeway for their quotation, especially when it comes to indirect and mixed quotes, there opens up a whole array of ways to manipulate the truth in subtle ways, and norms of fairness and transparency may be stretched. This was shown in the analysis of a number of examples, including the case of quotative wollen.

7 Relative lies 7.1 Introduction Participants of a discourse may be unsure with respect to the question whether an utterance was a lie. For instance, when asked by Barbie whether the weather was fine in Brighton, Ken answered “yes”. When Barbie checked the facts, she found that it was a sunny day except for a short period of extremely much rain. It is not easy to say whether she is right in accusing Ken of lying to her since this depends on the standards she applies to using the predicate “fine” with respect to weather. This simple case shows that lying may be relative to certain standards of assessment. In this chapter, we are going to explore some of these standards in more detail, drawing on recent contributions to relativism and disagreement. We will not be able to go deeply into the epistemological intricacies connected with this debate since our primary aim is to point out that matters of (un)truthfulness have linguistic repercussions, and that a realistic theory of lying must respect issues of “relative” lying, perspectivity, and disagreement.

7.2 Accuracy and disagreement Part of the ongoing debate on relativism, as opposed to contextualism, is the question how to account for disagreement in the case of predicates of taste. Suppose Ken utters, on the occasion of tasting wine together with Barbie, the speech act in (1): (1)

This Riesling is tasty.

Then he either acted truthfully because, according to his subjective standards of assessment, he likes the wine. Or he lied because, again according to his subjective standards of assessment, he dislikes the wine. Barbie, not exactly having access to Ken’s subjective standards of assessment, has only a very small chance at catching Ken in the act of lying, and this has to do with the vagueness of the predicate tasty. It is similar with other predicates of taste like delicious, funny, likely, etc. In a seminal article, MacFarlane (2007) elaborates the notion of disagreement. Contextualist approaches to disagreement usually assume that matters of taste are subjective. However, if they really are subjective, disagreement gets lost. The problem of “lost disagreement” may be solved better if a “scoreboard view”

212   

   Relative lies

is adopted. This view would require “a shared epistemic standard that changes as the conversation evolves” (MacFarlane 2007: 19). However, one drawback of the scoreboard view is that disagreement is restricted to a single conversation, hence “leaving inter-conversational disagreement unaccounted for” (MacFarlane 2007: 21). Two notions are important for a relativist approach, namely the speech acts of accepting and rejecting a proposition, and the accuracy of a proposition relative to a context. Accuracy is defined as in (2): (2)

Accuracy (MacFarlane 2007: 23) An acceptance (rejection) is accurate just in case the proposition accepted is true (false) at the circumstance of evaluation that is relevant to the assessment of the acceptance (rejection) in its context (or at all such circumstances, if there is more than one).

Here, “accurate” is a cognate of “true”, the latter predicate relating to sentences (propositions), the former to utterances (speech acts). The basics insight is that if two parties genuinely disagree, they assume that relative to a certain context C the acceptance and rejection of a certain proposition cannot both be accurate: (3)

Can’t both be accurate (relative to C) (MacFarlane 2007: 26) (a) There is a proposition that one party accepts and the other rejects, and (b) the acceptance and the rejection cannot both be accurate (as assessed from C).

In this definition, disagreement is made perspectival, or assessment-sensitive, because standards of taste are relative to disagreement in a certain context. Take the utterance Apples are delicious as containing the proposition on which Abe and Ben disagree. Be w the world in which Abe accepts and Ben rejects the proposition p = ‘that apples are delicious’, and be SCA the standard of taste of the assessor in the context of assessment CA: “According to Perspectival Accuracy, Abe’s acceptance is accurate (as assessed from CA) only if the proposition that apples are delicious is true at ‹w, SCA›, while Ben’s rejection is accurate (as assessed from CA) only if this proposition is false at ‹w, SCA›. […] Since no proposition can be both true and false at the same circumstance of evaluation, Abe’s acceptance of the proposition that apples are delicious and Ben’s rejection of his proposition cannot both be accurate. As assessed from CA, then, Abe and Ben disagree.” (MacFarlane 2007: 26–27)

Accuracy and disagreement   

   213

Elaborating on this scenario, we may ask whether lying with untruthful (dis) agreement is possible. We distinguish scenarios in which p, or q, or both contain the predicate of taste: (4)

a. b. c.

This wine is delicious. +> ‘You should buy it’. This wine is expensive. +> ‘It is delicious.’ This wine is dry. +> ‘It is delicious.’

Before embarking on the discussion of these cases, recall the simple case discussed in section 4.9 where the speaker was lying while simulating a particular standard taste, or simulating to have referred to a certain aspect under which their use of a predicate of taste had to be discussed. Suppose that in (4a), the speaker uses the predicate of taste untruthfully. Hence, their advice contained in the implicature is also contaminated. If the speaker acted truthfully, the implicature could also be deceptive, for instance, when the speaker knows that delicious wines are far too expensive. Suppose that in (4b), the speaker is truthful. Then the conversational implicature may nevertheless contribute to lying. After all, not all expensive wines are delicious. If the speaker was lying while using their untruthful implicature, the hearer has a certain chance to disagree on the truth of the conversational implicature. Taste predicates provoke acceptance or rejection, and this mechanism is basically the same with respect to p or q. Things are still getting trickier when considering (4c). In this case, two predicates of taste are involved. In order to successfully deceive the hearer, the hearer’s standards of taste should be known to the speaker. As a matter of fact, this turns out to be difficult. Standards of taste may be unstable, shaky, and moreover, may change during the conversation. From the speaker’s perspective, p may be true in (4c), but q may be false, for instance, when the speaker hates dry wine. In addition, p may be false and q may be true, for instance, when the speaker finds the wine sweet but still delicious for another reason. Finally, p and q both may be false. Thus, the speaker finds the wine sweet, and s/he does not find it delicious. What is the perspective of the hearer when they disagree with the speaker’s assessments? The main options are that the hearer has a sceptical attitude towards the speaker’s assessments (triggered by the use of the predicates of taste) both with respect to p and q, that the hearer rejects p, but accepts q, and vice versa, and finally, that they reject both p and q. Hence, we can safely conclude that there are relative implicatures:

214   

(5)

   Relative lies

Relative implicature An implicature is relative if its calculation is dependent on standards of taste as applied by the speaker and/or the hearer.

Lying by using untruthful relative implicatures is possible when the speaker successfully relies on stable standards of the hearer’s taste that are known to her/ him. Thus, the derivation of an implicature may depend on subjective standards of taste. Note in addition that a hearer may be mistaken in the derivation of an implicature; the implicature may not be intended by the speaker but the hearer may derive it on the basis of her/his subjective interpretation of a predicate of taste. This, then, is a case of misunderstanding. However, the point is that the hearer can never be sure how a predicate of taste used by the speaker was meant because s/he does not know the exact standard the speaker is applying. Thus, it is crucial how much the speaker and the hearer know about each other’s standards. Furthermore, standards of taste may (secretly) be switched. If they change during the discourse on the basis of arguments, persuasion, or accommodation, there is no lying involved but there arises the danger of misunderstanding or unclarity over the topic of the discourse. In sum, it is hard to lie for the speaker when s/he does not know the standards of taste of the hearer. Most safely s/he can only lie about her/his own assessments. MacFarlane (2007: 30) proffers a very appealing speculation about the functions of disagreement: “Perhaps the point is to bring about disagreement by leading our interlocutors into relevantly different contexts of assessments. If you say ‘skiing is fun’ and I contradict you, it is not because I think that the proposition you asserted is false as assessed by you in your current situation, with the affective attitudes you now have, but because I hope to change these attitudes. Perhaps, then, the point of using controversy-inducing assessment-sensitive vocabulary is to foster coordination of contexts.”

It goes without saying, then, that discourse on matters of taste may be used for deceptive purposes; lying about one’s own standards and covertly switching them with the purpose of deceiving the hearer are possible strategies of the liar. However, since the presupposition that the speaker is sincere is reduced from the hearer’s perspective, as the hearer knows that the use of predicates of taste invites to dissent and that s/he is free to accept or reject a proposition, there is also a chance that the liar is controlled in her/his enterprise.

Verum focus   

   215

7.3 Verum focus Predicates of taste are a class of verbal elements that open up an area of discourse in which deception appears easy because subjective standards of taste are hidden and cannot be checked with respect to truth or falsity. In this section, we are going to discuss a phenomenon where deception also appears hard to detect, but for another reason. The reason is that in this case the untruthful assertion is so strong that suspicions on the side of the hearer may be blocked, at least for a while. The phenomenon I am alluding to is called verum focus. In German, verum focus was first described by Höhle (1992). Höhle (1992: 112) analyzes verum focus as a kind of focus that emphasizes the manifestation of truth with respect to a thought (‘Hervorhebung der Bekundung der Wahrheit eines Gedankens’); the Fregean notion ‘thought’ refers to propositional content. The following mini-dialogue illustrates the phenomenon: (6)

Barbie:

Ken:

Ist es wirklich wahr, dass Ollie die Katze gefüttert hat? Is it really true that Ollie the cat fed has? ‘Is it really true that Ollie fed the cat?’ Ollie HAT die Katze gefüttert. Ollie HAS the cat fed. ‘It is true that Ollie fed the cat.’

As Lohnstein (2012) shows, verum focus is also to be seen with Yes/No-interrogatives and with wh-interrogatives (see 7), with the focus always being on the finite verb as in (6): (7)

a.

HAT Karl den Hund gefüttert? Has Karl the dog fed? ‘DID Karl feed the dog?’

b.

Wer HAT den Hund gefüttert? Who has the dog fed? ‘Who DID fed the dog?’

Moreover, in embedded dass-, w-, and d-sentences, verum focus may either be placed on the complementizer dass, w- and d-elements, or on the finite auxiliary as part of the complex verb (which is always in final position).

216   

(8)

(9)

   Relative lies

a.

Aber Emil glaubt, DASS Karl in Urlaub gefahren ist. But Emil believes that Karl in holiday gone is. ‘But Emil believes that Karl has gone on holiday.’

b.

Jetzt will ich wissen, WEN Karl eingeladen hat. Now want I know whom Karl invited has. ‘Now I want to know whom Karl has invited.

c.

Das ist der Wagen, DEN Karl gefahren hat. This is the car that Karl driven has. ‘This is the car Karl drove.’

a.

Aber Emil glaubt, dass Karl in Urlaub gefahren IST.

b.

Jetzt will ich wissen, wen Karl eingeladen HAT.

c.

Das ist der Wagen, den Karl gefahren HAT.

For the cases in (6) and (7) in which the verum focus is located in the finite verb, Höhle (1992: 114) finds that there is a meaning element VERUM attached to the verb such that this element is emphasized through the stress on the verb. The main problem is how to localize such a VERUM element in the syntactic and semantic structures of a sentence. Not only are different sentence types involved (declarative versus interrogative, root versus embedded sentences), the verum focus may also lie on lexical elements with distinct status, i.e., “complementizers” and finite verbs. Lohnstein (2012) convincingly argues that VERUM cannot be related to the finiteness of the verb. Properties contributing to finiteness, such as person, number, tense, and mood, may be contrasted, but this has nothing to do with the truth of the proposition expressed. Moreover, it does not seem plausible to relate truth to complementizers, w- and d-elements. Lohnstein’s approach is based on the hypotheses that verum focus has to do with focussing sentence modality, and that the effect of emphasizing the proposition’s truth (‘Wahrheitskundgabe’) may be derived on the assumption that verum focus reduces alternatives in the discourse situation. Thus, verum focus is portrayed as a sort of contrastive focus. In addition, Lohnstein argues that verum focus refers to sentence mood. Whereas illocutions are only associated with root sentences, sentence mood is also ascribed to embedded sentences. Because verum focus does not only occur in root sentences but also in embedded sentences, sentence modality must be the correct level of analysis. According to Lohnstein’s modular analysis, German sentences are mood phrases, i.e., projections of a head M0. Verum focus is located either in M0 (finite

Verum focus   

   217

verb, conjunction), or in SpM (the specifier of M’, where wh-elements are located, be it in root sentences or embedded sentences). These syntactic positions are associated with distinct semantic objects and sentence modalities. This is shown in the following table taken from Lohnstein: (10) SpM

Semantic objects and sentence modality (Lohnstein 2012) M0

Ø [–w]-XP [+w]-XP FIN0 Conj, Ø

semantic object

sentence modality

→ unmodified bipartition → reduced bipartition → differentiated bipartition

(yes/no-question) (judgement) (wh-question)

→ discourse anchoring → grammatical anchoring

(root) (embedded)

Put simply, reduced bipartition, as in judgements, amounts to positive or negative sentences. Only one of the two associated situations (p or -p) is acknowledged. Unmodified bipartition, as in yes/no-questions, leaves the judgement open, i.e., both alternatives are possible in the given situation. In wh-questions, we face a differentiated bipartition, i.e., all the alternative situations in which the wh-element is properly instantiated are possible; a reduction of these alternatives is left to the hearer. Let us turn to the discourse functions of verum focus. From (10) it follows that FIN0 is connected with discourse anchoring. As shown in (6), the discourse must contain a proposition whose truth is at stake. This is the case when Barbie asks whether it is really true that Ollie fed the cat. Alternatively, she may deny that Ollie fed the cat. In all-new contexts, verum focus is impossible. In verum focused wh-interrogatives and imperatives, the alternatives must be given in the context, for instance: (11)

Speaker: Weder Ken noch Barbie haben die Katze gefüttert. Neither Ken nor Barbie have the cat fed. ‘Neither Ken nor Barbie fed the cat.’ Hearer: Wer HAT denn nun die Katze gefüttert? Who HAS MP MP the cat fed? ‘Who did finally feed the cat?’

218   

(22)

   Relative lies

[Context: Ken walks around and is unsure which seat he should take.] Barbie: Nun NIMM dir endlich einen Stuhl! Now take youDAT finally a seat! ‘Now take a seat!’

Now Lohnstein argues contra Höhle that verum focus does not serve to emphasize the truth of a thought. Instead, the validity of truth (‘Wahrheitsgeltung’) comes about through the reduction of alternatives in the discourse. The reduction of alternatives has to do with a certain social status that enables the speaker to disrupt a discourse in an authoritarian way. Lohnstein’s main argument is that it is possible to lie by using verum focus. Imagine a situation in (6) in which Ken does know for sure that Ollie did not feed the cat. The effect induced by the verum focus is that his lie is stronger or more aggressive than without the verum focus. (It is similar with other devices like ich schwöre, ehrlich, ungelogen, etc.). Lohnstein presents the following example in order to flesh out his claim: (23)

[CEO Rainer Neske of the Deutsche Bank when taking over the Deutsche Postbank] Es WERden keine Stellen gestrichen. It AUX no jobs cancelled. ‘There are no cancellations of jobs.’

While the property of having the power certainly applies, a similar effect is available without any special social position. For instance, in (6) there is no special position of Ken necessary. What I propose is that the degree of commitment to the truth of p is what matters in assertions. Thus, using a verum focus, the speaker is committed to the truth to a higher degree than without the verum focus. This explains the power or aggressivity of the lie. The intuition that the use of verum focus may be invoked in cases in which someone aims at disrupting or ending a certain piece of discourse has to do with commitment. Thus, in the case of assertion, the speaker wants to express their strong commitment towards p and they want the hearer to take over this commitment. Similarly, in imperatives the speaker expresses their commitment to their aim that the hearer should make p true. In my view, then, verum focus is not so much a device that serves the reduction of alternatives, as Lohnstein argues, but a device that expresses the speaker’s strong commitment to p. This amounts to a powerful lie, in case the speaker acts untruthfully. This powerful lie has its risks. The liar risks that the evidence they can provide for the truth of their lie is under scrutiny. Hence, they should check in

The assertion of clarity   

   219

advance whether such evidence is available for the participants of the discourse. Only when evidence is unavailable or scarce could it be a promising move to use a verum focus.

7.4 The assertion of clarity In the last section, we considered verum focus, a phonological device connected to the act of highlighting the truth of a proposition. Occurrences of verum focus can be paraphrased with It is true that p. Not only can this construction be used in spoken and written language, there is also a similar construction, seen in (24b), that is interesting from a theoretical point of view and has led to a number of recent publications: (24)

a.

It is true that p.

b.

It is clear that p.

The main difference between asserting (24a) and simply asserting p is that in (24a), it is made explicit that p is true, while the assertion that the content p is true remains implicit in the simple assertion without the matrix sentence. Similarly, in explicit performative constructions like I promise to come the promise is made explicit, while in the utterance O.K. in the context of a question Will you come to my party? the promise is implicit. More difficult is the analysis of (24b) because clarity is not an in-built element of the definition of assertion. (The Griceans of course think of the Gricean supermaxim of Manner “Be perspicuous”.) Below we will be going into this issue. Since the matrix sentence is propositional, it can be negated: (25)

a.

It is not true that p.

b.

It is not clear that p.

Utterances like (25) may be used to explicitly deny the assertion that p is true or clear. In contrast, the negation of discourse adverbs like (26a, b) is not possible: (26)

a.

*Untruely, p.

b.

*Unclearly, p.

220   

   Relative lies

In these cases, negation is not possible, a fact that can easily be explained by the observation that lying cannot be made explicit. The same holds for the German discourse adverbs ehrlich gesagt und ungelogen: (27)

a.

Ehrlich gesagt, p. Sincerely said, p. ‘To tell you the truth, p.’ (‘To be absolutely honest, p.’)

b.

*Unehrlich gesagt, p.

c.

Ungelogen, p. Unlied, p. ‘Honestly, p.’

d.

*Gelogen, p.

The interesting point, however, is that all of these utterances, i.e., (24, 25, 27a, c), may be used for lying. Moreover, if they are used for lying, my intuition is that they are stronger lies than their unmodified counterparts. In this section, I will shortly go into the analysis of the assertion of clarity as proposed by Chris Barker and others (cf. Barker and Taranto 2003, Barker 2009, Barker 2011, Wolf and Cohen 2011). Since it is clear that I cannot deal with the relevant set of constructions in any detail, I will concentrate on the question how the it is clear that p construction may be used for lying. Two semantic theories dealing with the assertion of clarity are available, the belief theory (Barker and Taranto 2003, Wolf and Cohen 2011) and the justification theory (Barker 2009, Barker 2011). Here are the main tenets of the belief theory defined in (28): (28)

Belief theory of clarity (Barker and Taranto 2003) It is clear that p is true iff the degree of belief of the speaker and hearer in the truth of the proposition that p is greater than some vague standard.

Clear is considered as a vague predicate, just like predicates of personal taste. Now Barker (2009) argues that belief is neither necessary nor sufficient for the assertion of clarity: (29)

a.

It is reasonably clear that Mars is barren of life.

b.

It is clear that God exists.

The assertion of clarity   

   221

Both speaker and hearer may assent to (29a), yet believe that there is life on Mars. Furthermore, speakers and hearers may deny (29b), yet believe in God. Thus, belief is too subjective to account for clarity, and a justification theory of clarity, as proposed in Barker (2009), (2011), is called for. (30)

Justification theory of clarity (Barker 2011: 191) A proposition p is clear just in case all sufficiently normal worlds consistent with the common ground (i.e. consistent with publicly available evidence) are p-worlds. Then It is clear that p is true just in case the publicly available evidence justifies concluding that p.

Wolf and Cohen (2011) point out that the appeal of believing still remains attractive because of the distinctions between (a) simple clarity and personal clarity, and (b) adjectival clarity and adverbial clarity. While simple clarity refers to thirdperson assertions, personal clarity refers to personal perspectives as expressed in (31): (31)

It is clear to me/you/Johnny that Abby is a doctor.

Examples like (31) show that clarity is subjective, and that there may be disagreement on the reasons for clarity being invoked; this could have to do with the amount of knowledge publicly available, or certain standards of skepticism with regard to the predicate clear (p). However, Wolf and Cohen (2011: 168) present an example in which disagreement cannot be traced back to publicly available evidence: (32)

[Alex and Bill are jointly looking at a photograph of Abby wearing a stethoscope and smoking] Alex: It is clear to me that Abby is a doctor (because she is wearing a stethoscope). Bill: It is clear to me that she’s not (because she has an unhealthy habit).

As for the difference between adjectival clarity as in it is clear that p and adverbial clarity as in clearly, it can easily be shown that the two behave differently: (33)

a.

It’s unclear whether Abby is a doctor.

b.

*Unclearly, Abby is a doctor.

222   

(34)

   Relative lies

a.

If it’s clear that Abby will be a doctor then I, too, should apply to medical school.

b.

*If Abby will clearly be a doctor then I, too, should apply to medical school.

c.

It is clear that Abby is the best doctor in the hospital?

d.

??Is Abby clearly the best doctor in the hospital? (O.K. as an echoquestion)

Thus, the adverb of clarity cannot be negated (33) and is forbidden under the scope of conditional and interrogative operators (34). These differences, so Wolf and Cohen (2011) conclude, must have to do with belief because modal adverbs are related to the speaker’s belief, while the content in the adjectival case is expressed propositionally. Wolf and Cohen’s (2011: 179) proposal is “that the speaker is referring to the beliefs of people with sound judgment, people who are knowledgeable about the matter, who can reason rationally, and whose judgment is to be trusted.” This they call an “objectivised belief”. The speaker is reporting a belief, but this is an objectivized belief. If so, a paraphrase of (35a) would go like (35b): (35)

a.

It is clear that Abby is a doctor.

b.

It is believed by people with sound judgment etc. that Abby is a doctor.

But it is unclear to me whether this is the right way to go. First, Wolf and Cohen (2011: 179) hint at the inappropriateness of (36), showing that subjective belief nevertheless matters. (36)

#It is clear that Abby is a doctor, but I don’t believe it.

Second, while (35a) certainly suggests (35b) by way of implicature, this is not a correct paraphrase. This has to do with the rhetorical and persuasive effect of assertions like (35a) (that may be strengthened with totally, without any doubt, etc.; German: völlig/vollkommen klar, ohne jeden Zweifel klar, etc.), aiming at the hearer’s taking over the speaker’s belief that p. Thus, it is clear that backing an assertion by public authority is a good reason to believe something if one is unsure about what to believe. It follows that such utterances are very well suited for lying. If Ken does not believe that Abby is a doctor but asserts (35a), he is lying. This is a tricky

The assertion of certainty   

   223

lie indeed since the justification can be shifted to public opinion (which is not easily to be proven). Barker (2011: 202) mentions a “side effect” of the assertion of clarity, namely “putting the truth of a proposition beyond negotiation”. According to him, this is so conventional that one can even begin a conversation with the assertion of clarity. Again, this shows that lying while asserting clarity is easy, yet cannot easily be detected. As for the correct theory, Barker (2011) also points to the fact that it is hard to explain what these objectivized beliefs really are. As for personal clarity, he argues that “on the justification approach, then, the difference between personal clarity and simple clarity is that personal clarity explicitly indicates who must have (conscious!) access to the relevant evidence” (Barker 2011: 194), and with regard to the distinction between adverbial and adjectival clarity, he states that “apparently, a simple belief theory for adverbial clarity faces the same problems it faces for adjectival clarity” (Barker 2011: 196). In sum, then, I would like to argue that if a speaker of (35a) is asked what the reasons for their assertion are, they should be able to justify their assertion (in accordance with the maxim of Quality.). Hence, it turns out that both belief and justification play a role in the assertion of clarity. Our definition of assertion requires that the speaker believes p, and that they should be in the possession of evidence for p. Both aspects are also relevant for the assertion of clarity, simply because it remains an assertion. When it comes to adverbial and personal clarity, it is, as in the case of subjective attitudes on taste, hard to verify or falsify whether the speaker said the truth. Nevertheless, they are able to lie.

7.5 The assertion of certainty If someone utters it is clear that p, they suggest that they are certain that p. The degree of certainty a speaker has with regard to an asserted proposition is relevant for lying. To spell out this intuition requires an analysis of certainty. I will assume that being certain, like being clear, has to do with the justification of a speaker’s belief that p. Thus, a certain belief that p is certain when it is justified for a speaker in the highest degree. Chisholm (1976, 1989) has defended definitions of certainty that are based on these assumptions. Here I stick to the version in Chisholm (1976), but replace the variables (his h = p, his i = q): (37)

Certainty (Chisholm 1976: 27) p is certain for S at t =df (i) Accepting p is more reasonable for S at t than withholding p (i.e., not accepting p and not accepting not-p) and (ii) there is no q such that accepting q is more reasonable for S at t than accepting p.

224   

   Relative lies

As Reed (2008: 6) explains, clause (i) ensures that the speaker S has “some measure of positive justification” for p, because “if she had no justification for it, it would be more reasonable for her to withhold” with respect to p, i.e., being ignorant. So if Ken believes that Barbie has a Ph.D., he should have some evidence for this belief; if he did not, he should withhold p. Clause (ii) says “that those beliefs of the subject are certain which are at the highest levels of justification for her” (Reed 2008: 6). For instance, if Ken has seen Barbie’s Ph.D. diploma, then he will be most certain that his belief is justified. With respect to (37), Reed (2008) points out several problems. One problem is that p “is the most highly justified belief the subject [= the speaker] has, but it is still not very highly justified (e.g., it may not even be sufficiently justified to count as knowledge).” However, this is a welcomed feature of the definition because speakers may indeed hold very weakly justified beliefs, yet tend to assert p. Note the recently coined notion of truthiness that is traced back by Narvaez (2010) to the comedian Stephen Colbert; Colbert is reported to have used this notion for the first time in his TV show “The Colbert Report” on October 17th, 2005: (38)

Truthiness Things that a person claims to know intuitively or “from the gut” without regard to evidence, logic, intellectual examination, or facts.

Thus truthiness can be defined as an epistemic state in which p is asserted with certainty although there is no justification for p available. (Note that this plays a role in certain varieties of bullshit.) Another worry with the definition in (37) is, according to Reed (2008: 7), that there the following problematic scenario can arise: “Two subjects each believe that p, and in each case the belief is justified to degree n. For the first subject, the belief counts as certain because none of her other beliefs have a higher level of justification. But, for the second subject, the belief in question is not certain because she does have another belief that is slightly more justified. If certainty really is grounded in epistemic justification, though, this should not be possible. If a given justification makes a belief certain for one subject, it should do so for everyone.”

However, this scenario describes a situation we already dealt with, namely the dispute about predicates of taste (see section 7.2). So we admit that justification may not be given in an objective, but only in a subjective way. What is still lacking in the above definition is an idea about how a speaker could become more or less certain. I propose the following definition (see section 5.6):

Reliability and selfless assertion   

(39)

   225

Degrees of Certainty The speaker is at t certain that p to a high degree, iff (i) the addition of q to their knowledge would not make them more certain, and (ii) the addition of r to their knowledge would not make them less certain.

The point is, then, that the speaker is in a stable position of certainty with respect to p. This does not exclude that there is a possibility of becoming more or less certain if the knowledge base is changed. For the sake of illustration, consider the case already presented in section 5.6, here repeated for convenience. Suppose that Ken asserts that Barbie has a Ph.D. (= p). Now Ben tells Ken that he had seen a doctoral cap in Barbie’s apartment. But this does not make Ken more certain. Although his assumption is supported by Ben’s information, he does not change his mind. This would only happen if he was in doubt with respect to p. If Ken however asserted that Barbie had not visited college (= r), Ken could become less certain about his belief. He knows that having visited college is a precondition for having a Ph.D. Now Ken is very certain about his knowledge (he is stubborn). He thinks that even if the information given by Ben is true, there must be a possibility to do a doctor’s degree without having visited a college. So Ken blocks his becoming less certain. In a way, then, someone is lying when they assert something (i) expressing certainty when they are uncertain, or expressing uncertainty when they are certain, or (ii) expressing certainty or uncertainty to a higher degree than being adequate with respect to their knowledge base. For instance, Ken is uncertain about Barbie’s degree but nevertheless expresses his certainty that Barbie holds a Ph.D. Or he states that it is absolutely certain that Barbie holds a Ph.D. when he has only a vague assumption that this could be the case (because Barbie often quotes Shakespeare).

7.6 Reliability and selfless assertion In the foregoing section, I have touched upon norms of assertion such as the certainty norm (see also section 5.6). In this section, I am going to focus on yet another facet of this issue. Lackey (2008) attacks two widespread assumptions about the connection of assertion and knowledge. The first thesis is the thesis of Transmission of Epistemic Properties (TEP), the second is the Knowledge Norm of Assertion (KNA).

226   

   Relative lies

Let us start with a short discussion of the TEP. There is a necessity and a sufficiency thesis with regard to the TEP; the necessity thesis is seen below. (40)

Transmission of Epistemic Properties – Necessity (Lackey 2008: 39) For every speaker, A, and hearer, B, B knows (believes with justification/warrant) that p on the basis of A’s testimony that p only if A knows (believes with justification/warrant) that p.

So B can only know p on the basis of A’s testimony (an assertion) if A knows p. We can add that in the case of A’s lying, B will know a false content of A’s testimony without the liar A’s knowing this content. The TEP-Sufficiency Thesis (e.g., held by Adler 2006) is seen in (41): (41)

Transmission of Epistemic Properties – Sufficiency (Lackey 2008: 39) For every speaker, A, and hearer, B, if (1) A knows (believes with justification/warrant) that p, (2) B comes to believe that p on the basis of the content of A’s testimony that p, and (3) B has no undefeated defeaters for believing that p, then B knows (believes with justification/warrant) that p.

The parlance of “no undefeated defeaters” calls for an explanation. Lackey (2008) draws a distinction between psychological defeaters on the one hand, and normative defeaters on the other hand. A psychological defeater “is a doubt or belief that is had by S and that indicates that S’s belief that p is either false or unreliably formed or sustained” (Lackey 2008: 44). A normative defeater “is a doubt or belief that S ought to have and that indicates that S’s belief that p is either false or unreliably formed or sustained” (Lackey 2008: 45). More generally, defeaters “contribute epistemically unacceptable irrationality to doxastic systems and, accordingly, knowledge (justification/warrant) can be defeated or undermined by their presence” (Lackey 2008: 45). For Lackey, both (40) and (41) are completely misguided. Directed against the TEP in both variants, she develops several stories (thought experiments) that are intended to show that it is flawed. One of the set of stories directed against (40)  – the TEP-Necessity  – is called CREATIONIST TEACHER. I quote it here in full length:

Reliability and selfless assertion   

(42)

   227

CREATIONIST TEACHER (Lackey 2008: 48) Stella is a devoutly Christian fourth-grade teacher, and her religious beliefs are grounded in a deep faith that she has had since she was a very young child. Part of this faith includes a belief in creationism and, accordingly, a belief in the falsity of evolutionary theory. Despite this, she fully recognizes that there is an overwhelming amount of scientific evidence against both of these beliefs. Indeed, she readily admits that she is not basing her own commitment to creationism on evidence at all but, rather, on the personal faith that she has in an all-powerful Creator. Because of this, Stella does not think that religion is something that she should impose on those around her, and this is especially true with respect to her fourth-grade students. Instead, she regards as her duty as a teacher to involve presenting material that is best supported by the available evidence, which clearly includes the truth of evolutionary theory. As a result, after consulting reliable sources in the library and developing reliable lecture notes, Stella asserts to her students, “Modern-day Homo sapiens evolved from Homo erectus,” while presenting her biology lesson today. Though Stella herself neither believes nor knows this proposition, she never shares her own personal faith-based view with her students, and so they form the corresponding true belief solely on the basis of her reliable testimony.

The upshot of this story is to argue for the following relationship between reliability and knowledge: “What CREATIONIST TEACHER reveals is that an unreliable believer may nonetheless be a reliable testifier, and so may reliably convey knowledge (justified/warranted belief) to a hearer despite the fact that she fails to possess it herself.” (Lackey 2008: 41) However, since Stella obviously presents something as true which she does not believe in, I would argue that she is lying. For Lackey, Stella’s utterance is a case of a so-called “selfless assertion”. As for “selflessness” and the issue of Stella’s reliability, I would like to argue that it is a special case of lying. First, it could be a case of self-deception. Stella has two conflicting states of mind. In her private, religious world, she believes that an all-mighty Creator created the world, while in her professional, social world she believes that homo sapiens developed from homo erectus. The contents are in conflict, yet Stella is not aware of this and she fluently switches between these two worlds. Second, Stella’s utterance is a lie (with respect to her belief system); however, it is a prosocial lie because Stella considers the majority opinion and does not want to infringe on this. Indeed, she would not impose her private religious

228   

   Relative lies

feelings on others since group harmony is also a value that she sticks too. So her utterance is meant to stabilize the community instead of confronting it with her religious beliefs. She of course knows that her beliefs are the right ones and would ultimately be justified, but in order not to violate the beliefs of the majority, she accommodates herself to it. Third, her utterance is bullshit. In a way, she has a lack of concern for the truth and she does not want the students to become aware of this. Moreover, she presents the evolutionist evidence with far more certainty than would be apt with respect to her religious feelings. So here we would have a case in which one speaks the truth (supposing that evolutionists are right), yet is bullshitting with regard to their true beliefs. Lackey (2008) presents several other stories (CONSISTENT LIAR, PERSISTENT BELIEVER, SERIOUS STUDENT) directed against TEP-Necessity we cannot go into here. Let us turn to her arguments against (41), TEP-Sufficiency, that are again presented in the form of stories/thought experiments. Here, I would like to pick out COMPULSIVELY TRUSTING and ALMOST A LIAR (the other being A LUCKY CHOICE, p. 68). COMPULSIVELY TRUSTING is about Bill who has an obsessively romantic interest in Jill, and therefore compulsively trusts in Jill’s words. So when Jill asserts she has seen an orca whale, he believes this. The point of this story, according to Lackey (2008: 67), is “that while a speaker may be both a perfectly reliable believer and testifier, a hearer may be so constituted as to prevent the epistemic properties of a speaker’s belief from being transmitted to her.” In a way, Bill is an insensitive addressee for he believes everything Jill asserts. A variant of this story is ALMOST A LIAR:

Reliability and selfless assertion   

(43)

   229

ALMOST A LIAR Phil is compulsively-trusting-Bill’s twin brother, though he is not himself compulsively trusting with respect to Jill’s testimony. In fact, he is, from an epistemic point of view, quite healthy: he trusts those whom he has good reason to trust – or at least those whom he has no clear reason to distrust – and distrusts those whom he has good reason to distrust. Yesterday, while taking his afternoon walk, Phil ran into Jill just after Bill did, and she also told him that she had seen an orca whale while boating earlier that day. Phil, having acquired very good reasons for trusting Jill over the five years he had known her, readily accepted her testimony. It turns out that Jill did in fact see an orca whale on the boat trip in question, that she is very reliable with respect to her epistemic practices, both in general and in this particular instance, that she is generally a very reliable testifier, and that Phil has no reason to doubt the proffered testimony. However, in order to promote a whale watching business she is in the process of starting, she would have reported to Phil – in precisely the same manner – that she had seen an orca whale even if she hadn’t. (Of course, she wouldn’t have believed that she had seen an orca whale if she hadn’t.) Moreover, given the pattern of the whales’ travel combined with the particular time of year it is, it is in fact quite surprising that Jill saw an orca whale when and where she did.

So Jill does not lie, but would have lied in an alternative world, and epistemically “healthy” Phil is justified in trusting Jill’s assertion. Yet Lackey (2008: 70) concludes from her example: “Otherwise put, in nearby worlds where it is false that an orca whale was in the relevant body of water without this being so, Phil believes that there was one, and there are nearby worlds where Phil believes that an orca whale was in the relevant body of water without being so. Given this, the true belief that Phil forms on the basis of Jill’s testimony not only fails to qualify as knowledge, it also fails to be justified or warranted.”

Here, I do not understand Phil’s relation to an alternative world. When interpreting the story in (43), I would conclude that Phil is epistemically justified to believe what Jill asserted, independent of an alternative situation in which she lied. Even if she lied, he would be epistemically justified in assuming that the content of her lie was true (after all, this is how lies function). Therefore, I am not convinced that the case ALMOST A LIAR is counterevidence for TEP-Necessity. As mentioned before, Lackey argues does not only argue against the TEP, but also against the Knowledge Norm of Assertion (KNA). The idea is that TEP-N cannot be supported by KNA because KNA is wrong.

230   

(44)

   Relative lies

Knowledge Norm of Assertion (Lackey 2008: 103) One should assert that p only if one knows that p.

According to Lackey, the KNA is wrong. Instead, she proposes the Reasonable to Believe Norm of Assertion (RTBNA) (Lackey 2008: 125): (45)

Reasonable to Believe Norm of Assertion (RTBNA) One should assert that p only if (i) it is reasonable for one to believe that p, and (ii) if one asserted that p, one would assert p at least in part because it is reasonable for one to believe that p.

Counterexamples to (44) are of the type “selfless assertion”. The story CREATIONIST TEACHER already contained such a selfless assertion. Beside the stories MODIFIED CREATIONIST TEACHER and DISTRAUGHT DOCTOR, she presents the story of the RACIST JUROR (Lackey 2008: 110): (46)

RACIST JUROR Martin was raised by racist parents in a very small-minded community and, for most of his life he shared the majority of beliefs held by his friends and family members. After graduating from high school, he started taking classes at a local community college and soon began recognizing some of the causes and consequences of racism. During this time, Martin was called to serve on the jury of a case involving a black man on trial for raping a white woman. After hearing the relatively flimsy evidence presented by the prosecution and the strong exculpatory evidence offered by the defense, Martin is able to recognize that the evidence clearly does not support the conclusion that the defendant committed the crime of which he is accused. In spite of this, however, he can’t shake the feeling that the man on trial is guilty of raping the woman in question. Upon further reflection, Martin begins to suspect that such a feeling is grounded in the racism that he still harbors, and so he concludes that even if he can’t quite come to believe that the defendant is innocent himself, he nonetheless has an obligation to present the case to others this way. Shortly after leaving the courthouse, Martin bumps into a childhood friend who asks him whether the “guy did it.” Despite the fact that he does not believe, and hence does not know, that the defendant in question is innocent, Martin asserts, “No, the guy did not rape her.”

Reliability and selfless assertion   

   231

This example, as the other examples, shall demonstrate “that it is a mistake to require proper assertion to pass through the doxastic states of the asserter.” (Lackey 2008: 113) Again, I would argue that this utterance was a lie since Martin asserted something he did not believe in and thus deceived the hearer about his true opinion. As for the possible motives of his lying, there are the three options I have already presented with respect to the case of the CREATIONIST TEACHER, i.e., self-deception, prosocial lie, and bullshitting. Lackey is well aware of the possible move (I adhere to) that her counterexamples for the KNA are not suited for her purpose because they contain lies. She presents three arguments against her examples and rebuts them in turn. According to the first argument, her cases of selfless assertion “do not qualify as genuine assertions” (Lackey 2008: 114) because they show a lack of belief on the part of the speaker. Lackey replies that having a certain belief is not necessary for assertion since “to assert a lie” (Lackey 2008: 115) is also possible, lying typically lacking the belief that p is true. However, “to assert a lie” is a misconception: lying is insincere assertion, but it is impossible to assert that one is lying. The second argument against Lackeys approach, as reported by Lackey (2008: 115), is that the crucial utterances are lies. Indeed, this is also my view. Yet Lackey argues that there is no intent to deceive on the part of the speaker. In her own words (Lackey 2008: 116): “(…) notice first that in cases of selfless assertion, there is absolutely no intention on the part of the asserter to deceive or otherwise to mislead. Indeed, quite the contrary is true – the asserter in question positively intends to not deceive or mislead the hearer and, as a result, asserts what she herself does not believe.” However, I do not find this convincing since the speaker clearly deceives the hearer about their true attitudes. In order to see that these attitudes matter, a change of context is sufficient. For instance, the creationist teacher does not only teach biology, but also religion. Or the creationist teacher gives courses in her religious, creationist community. So the asserter does not want to deceive her audience about the majority opinion on evolutionism, but she wants the audience to be deceived about her beliefs and hence is not truthful. The third objection against Lackeys proposal is a sort of test. If the asserter were asked “whether he or she believes or knows the proposition asserted”, and the answer were no, the questioner would “feel cheated or deceived upon learning that the answer to these question is no” (Lackey 2008: 117). Lackey does not agree: “For instance, surely the childhood friend of Martin [in RACIST JUROR, JM] would not prefer that he assert what he truly believes about the defender’s guilt, for this would result in Martin offering an assertion based purely on racism.” But I think the question is not what the childhood friend would prefer but what the truthful opinion of Martin is. Suppose that the childhood friend is one of Martin’s

232   

   Relative lies

family neighbors, small-minded like the whole neighborhood. Then, of course he would be interested in Martin’s true opinion, or in the question how much Martin is still sticking to his old persuasions, how much the community college had influenced him, etc. Lackey (2008: 118) maintains that “it should be widely accepted that such cases of selfless assertion provide convincing counterexamples to the KNA” – I would object that this cannot be the case, for these cases are “selfless lies” at best. In my view, the students of the creationist teacher as well as Martin’s childhood friend trust their interlocutors and find them a reliable source of knowledge, yet are duped by them about their true beliefs. The idea of proposing the Reasonable to Believe Norm of Assertion (RTBNA) is to develop a weaker norm than the KNA, one that is compatible with the stories proffered against it. The asserter respecting a Reasonable Belief Norm of Assertion (this is a precursor of the fleshed-out RTBNA in 45) “is not subject to criticisms for asserting that which she reasonably believes she knows, even if infelicitous circumstances lead her to assert a proposition that as a matter of fact falls short of her knowledge” (Lackey 2008: 108). Not only do some of Lackey’s examples fall short of fitting to the RTBNA, as she acknowledges in a footnote (Lackey 2008: 106, Fn. 8), she also proposes a further norm dubbed NMNA: (47)

Not Misleading Norm of Assertion (Lackey 2008: 134) S should assert that p in context C only if it is not reasonable for S to believe that the assertion that p will be misleading in C.

This looks like a Gricean maxim, as Lackey (2008: 135) rightly observes. She proposes that it might be akin to the maxim of Quantity, because the selfless asserters may be liable to not having told the whole story of their beliefs. However, if taken as a vehicle to soften the inadequate effects of selfless assertions, the NMNA seems superfluous to me since there is a maxim of Quality or a sincerity condition regulating proper assertion.

7.7 Conclusions In this chapter, we have shown that lies may be relative, i.e., they have to do with subtle matters of accuracy, clarity, and certainty. When it comes to disagreement, lying is very easy since the risk of the liar of being caught in the act of lying is minimal. Also, it is hard to detect the liar when standards of accuracy, clarity, and certainty are hidden from direct observation and may be shifted during the conversation according to the liar’s overall goal to deceive their addressee. To

Conclusions   

   233

set a standard or norm for proper assertion is by no means easy. The widespread idea of a knowledge-norm for assertion is attacked by Lackey by her presenting cases of so-called selfless assertions. Here, she argues, knowledge is transmitted without the asserter believing what they assert; yet these utterances, according to Lackey, are no lies. While the alternative model of Lackey, the Reasonable to Believe Norm of Assertion (RTBNA) is debatable, I have objected that cases of selfless assertions are lies.

Epilogue The main idea behind this book was that a comprehensive notion of lying should include lying by deliberately using false implicatures and false presuppositions. Implicatures and presuppositions are additional propositions that are derived from an utterance in an interactional context. I take it that these additional propositions are intended in the same way as the illocutionary point of the triggering utterance. In the process of understanding the speaker, the hearer’s task is to find out what the total speaker meaning is. It is clear that implicatures and presuppositions are a proper part of the total speaker meaning (or the total signification of the speaker’s utterance, as Grice 1989b: 41 puts it). From a linguistic perspective, we need a view of the grammar-pragmatics interaction in general and of the semantics-pragmatics interaction in particular. As background models, we considered models like the minimalist (Gricean) approach to the semantics-pragmatics divide and the contextualist (Neogricean) approach by Levinson. In addition, we touched upon frameworks by Relevance theorist Robyn Carston and Kent Bach. In Garrett and Harnish (2007: 68), a nice diagram visualizes different “pragmatic options” that are chosen by Grice, Sperber and Wilson, Carston, Recanati, and Bach by trying to model the path from linguistic meaning to conversational implicature (see also Table 3, section 2.6, showing Levinson’s model). According to this diagram, there are two boxes that are not disputed and are somehow taken for granted, namely Pragmatics-1 and Pragmatics-2. Pragmatics-1 contains “mechanisms of contextual disambiguation”. These mechanisms have as their output an “operative (applied timeless) meaning” which are input to “mechanisms of reference fixing”. Pragmatics-2 simply contains mechanisms of conversational implicature, having as their output (conversational) implicature. The pragmatic dispute is over the processes that lead from Pragmatics-1 to Pragmatics-2: impliciture, explicature, mechanisms of saturation, and the role of “what is said”. However, this is still a very rough picture since it neglects the role of the grammar and lexicon, as well as the role of the speech act and discourse. Moreover, it is unclear where truth conditions are to be located. What the “right” conception of the semantics-pragmatics interface is, is of course open to dispute. Even the most recent linguistic attempts at giving a comprehensive view on what pragmatics is (cf. Ariel 2010, Cummings 2009, Allan and Jaszczolt 2012, Huang 2012) offer a wealth of appealing, yet competing approaches. In this situation, I take it that all works are welcome that analyze linguistic phenomena at the semantics-pragmatics interface in more detail, and this is what I have tried with respect to lying. Enlarging this broad picture of the semantics-pragmatics interface with a look at lying, one can point out how decisions of the speaker as well as avail-

Epilogue   

   235

able mechanisms lead to different pragmatic options. For instance, the speaker chooses an adequate sentence type, e.g., a declarative sentence and they choose appropriate intonation patterns and accents (e.g., verum focus). In addition, s/he chooses lexical items fitting to this purpose; e.g., no hesitation markers but sentence adverbs like honestly, predicates of taste, etc. This leads to a proposition that is accompanied by certain attitudes (sentence mood, modal particles, etc.). This proposition may still be underdetermined, hence it will be enriched (explicature/impliciture). The enriched proposition is then evaluated with respect to truth-conditions. Given a certain context, a direct or indirect speech act results. For instance, the utterance of the declarative sentence is understood as an assertion. On the basis of certain speech acts, conversational implicatures may be triggered. The single speech act is embedded in the discourse and takes part in determining the common ground. The strategy chosen by the speaker may widely differ according to different contexts. The important thing is that lying is, on the one hand, an act which must be carefully planned by the liar, yet, on the other hand, the liar is free to choose how her/his plan is carried out in an optimal way. Most philosophical analyses of lying are not very deeply founded in linguistics. Typically, matters of linguistic structure are neglected. There are exceptions to this rule, e.g. when Sorensen (2012) reflects about the possibility of lying by using conditionals, but in most cases, this does not go very far. That assertions have to do with declarative sentences, with lies typically being insincere assertions, is seldom reflected. Whether it is possible to lie with exclamatives or other sentential moods is left in the dark, because one does not know to which extent an utterer of an exclamation is asserting something. How intentionally false implicatures and presuppositions may be related to different sentence types and illocutionary acts is not a topic of a typical analysis in the philosophy of language. In this book, I tried to compensate for this lack of linguistic analysis. In a lucid analysis, Bach (2013) warns against the “lure of linguistification”. By this, he understands “attributing linguistic properties to non-linguistic phenomena”, and also “attributing non-linguistic properties to linguistic items, treating these properties as if they were linguistic” (Bach 2013: 1). One typical fallacy is the so-called intuition fallacy (Bach 2013: 7): “If a certain proposition is the ‘intuitive truth-conditional content’ of an utterance, then it is the semantic content of the uttered sentence.” An instance of this fallacy, Bach explains, is seen in the analysis of the sentence Gene hasn’t gone to work as having the proposition Gene hasn’t gone to work today as its “intuitive truth-conditional content”. This is a classical example demonstrating underdetermination and the need for enrichment. However, as Bach observes, the idea that “the time restriction is part of the semantic content” has to do with our imagination of a typical context of an utterance (Bach 2013: 8): “So our supposedly semantic intuition is insensitive to

236   

   Epilogue

the differences between the semantic content of the sentence and what it is most likely to convey.” Most importantly, “there is nothing in the sentence to impose that [time] restriction”. In a similar vein, Lasersohn (2012) warns against contextualist attacks against compositionality. While there certainly is something like the “lure of linguistification”, there might be also be something like the “lure of pragmatification” or the “lure of philosophication”, for instance, when different linguistic phenomena are mapped on very broad philosophical notions. This can lead to somewhat premature generalizations like “Ironical utterances are always echoic in nature”, or “All utterances strive for maximal relevance”, “echo” and “relevance” being understood as genuine pragmatic or philosophical concepts. I do not deny that pragmatification or philosophication might constitute a fruitful (pragmatic) research strategy. If we take “bullshit” to be a philosophical concept, then it is tempting to ask to which extent this concept is also a linguistic or pragmatic concept. However, when talking about the proper semantics-pragmatics divide, there is a lure of linguistification as well as a lure of pragmatification/philosophication. There are typical notions that mirror the lures lurking around the dividing lines. A case in point is the notion of context. In Meibauer (2012b), I tried to show that not much is known about what a context “really” is. We have a received set of ideas on what aspects a context contains, but when it comes to competing approaches, e.g., minimalism versus contextualism, scholars tend to pick out those elements from the set that best suits their particular approaches. Nothing is wrong about such strategies, although one should bear in mind that there are further strategies for fruitful research. What I allude to, then, is experimental pragmatics. For instance, the notion of ‘what is said’ appears to be notoriously unclear. It is not easy to say whether this notion is needed in addition to the more traditional notion of “propositional content” (Terkourafi 2010). Thus it is useful to test intuitions about ‘what is said’ directly, as Doran et al. (2012) did. Note that this is a way to go with respect to the concept of lying, too. The seminal approach of Coleman and Kay (1981), planned as a study in prototype semantics, was as well a study in experimental pragmatics. I hope that there will be further experimental studies like the ones by Hardin (2010) and Arico and Fallis (2013). After all, we want to know what the concept of lying looks like for real speakers and hearers. There is a growing interest in (inter)cultural pragmatics of lying, the crucial question being whether all liars are alike or whether there are culturally parameterized liars (cf. Vincent Marelli 2004, Blum 2007). Again, what we need is both a comprehensive linguistic theory of lying and an understanding of how cultural and social aspects of lying fit into this theory. In order to gain insights in this respect, more experimental work will be helpful.

Epilogue   

   237

A typical, well-respected method in the philosophy of language is to present a series of definitions of lying and their related approaches to then point out their weaknesses, etc. (e.g., Carson 2010, Saul 2012). Mostly, this is a very explicit method of arguing. However, one should bear in mind that examples typically are on speech-act level. Thus, we should not forget that lying happens within discourses. Another typical and dignified method is to present stories about lying, including more or less funny scenarios (e.g., the stories invented by Carson 2010, Lackey 2008). One should bear in mind that quite often the results of interpreting these stories are dependent on intuitions and the weight one gives to single aspects of the respective scenario (cf. Meibauer 2012a). In recent approaches to lying, e.g., Fallis (2012), there is a tendency to analyze lying from one particular angle, in this case the maxim of Quality. While such an approach may be very fruitful for informing the philosophical dispute, I tried to develop a broader perspective on lying, i.e., one that integrates insights from a number of linguistic, philosophical, and psychological areas. This synthesis is far from being perfect, yet seems promising with respect to its overall architecture. I hope that future studies on the semantics-pragmatics interface can profit from it.

References Abbott, Barbara. 2000. Presuppositions as nonassertions. Journal of Pragmatics 32: 1419–1437. Abbott, Barbara. 2003. Some Notes on Quotation. In Hybrid Quotations (Belgian Journal of Linguistics 17), Philippe de Brabanter (ed.), 13–26. Amsterdam, Philadelphia: Benjamins. Abels, Klaus. 2010. Factivity in exclamatives is a presupposition. Studia Linguistica 64.1: 141–157. Adenzato, M. and Ardito, R.B. 1999. The role of theory of mind and deontic reasoning in the evolution of deception. In Proc. 21st Conf. Cogn. Sci. Soc., M. Hahn and S.C. Stoness (eds.), 7–12. Mahwah, NJ: Erlbaum. Adler, Jonathan E. 1997. Lying, deceiving, or falsely implicating. The Journal of Philosophy XCIV: 435–452. Adler, Jonathan E. 2006. Epistemological Problems of Testimony. Stanford Encyclopedia of Philosophy Allan, Keith and Kasia M. Jaszczolt (eds.). 2012. The Cambridge Handbook of Pragmatics. Cambridge: Cambridge University Press. Alston, William P. 2000. Illocutionary Acts and Sentence Meaning. Ithaca, London: Cornell University Press. Altmann, Hans. 1993. Satzmodus. In Syntax. Ein internationales Handbuch zeitgenössischer Forschung, Bd. 1, Joachim Jacobs, Arnim von Stechow, Wolfgang Sternefeld und Theo Vennemann (eds.), 1006–1029. Berlin, New York: De Gruyter. Amaral, Patricia, Craige Roberts, and E. Allyn Smith. 2007. Review of The Logic of Conventional Implicatures. Linguistics and Philosophy 30, 707–749. Arico, Adam J. and Don Fallis. 2013. Lies, Damned Lies, and Statistics: An Empirical Investigation of the Concept of Lying. Philosophical Psychology 26 (6): 790–816. Ariel, Mira. 2010. Defining Pragmatics. Cambridge: Cambridge University Press. Atlas, Jay David. 2005. Logic, Meaning, and Conversation. Semantic Underdeterminacy, Implicature, and Their Interface. Oxford: Oxford University Press. Austin, John L. 1962. How to do things with words. Cambridge, Mass.: Harvard University Press. Autenrieth, Tanja. 1997. Tautologien sind Tautologien. In: Pragmatik. Implikaturen und Sprechakte. (Linguistische Berichte, Sonderheft 8), Eckard Rolf (ed.), 12–32. Opladen: Westdeutscher Verlag. Bach, Kent. 1975. Performatives are statements too. Philosophical studies 28: 229–236. Bach, Kent. 1999. The Myth of Conventional Implicature. Linguistics and Philosophy 22: 327–366. Bach, Kent. 2002. Seemingly semantic intuitions. In Meaning and Truth, Keim J. Campbell, M. O’Rourke, and D. Shier (eds.), 21–33. New York: Seven Bridges. Bach, Kent. 2006. The top 10 misconceptions about implicature. In Drawing the Boundaries of Meaning. Neo-Gricean studies in pragmatics and semantics in honor of Laurence Horn, Betty J. Birner and Gregory Ward (eds.), 21–30. Amsterdam, Philadelphia: Benjamins. Bach, Kent. 2010. Impliciture vs Explicature: What’s the Difference? In Explicit Communication. Robyn Carston’s Pragmatics, Belén Soria and Esther Romero (eds.), 126–137. Basingstoke: Palgrave Macmillan. Bach, Kent. 2013. The Lure of Linguistification. In What is said and What is not, Carlo Penco and Filippo Domaneschi (eds.), 1–11. Stanford: CSLI Publications.

References   

   239

Bach, Kent and Robert M. Harnish. 1979. Linguistic Communication and Speech Acts. Cambridge, Mass., London: The MIT Press. Bach, Kent and Robert M. Harnish. 1992. How performatives really work: A reply to Searle. Linguistics and Philosophy 15: 93–110. Barker, Chris. 2009. Clarity and the grammar of scepticism. Mind and Language 24(3): 253–273. Barker, Chris. 2011. Commentary on Wolf and Cohen: Reasoning about Public Evidence. In Vagueness and Language Use, Paul Égré and Nathan Klinedinst (eds.), 191–203. Basingstoke: Palgrave Macmillan. Barker, Chris. 2013. Negotiating Taste. Inquiry 56: 240–257. Barker, Chris and Gina Taranto. 2003. The Paradox of Asserting Clarity. In Koskinen, P. (ed.), Proceedings of the Western Conference on Linguistics (WECOL) 2002, Vol. 14, 10–21. Department of Linguistics, California State University, Fresno. [http://semanticsarchive. net/Archive/ TY0OWM1Z/] Barker, Stephen. 2004. Renewing Meaning. Oxford: Oxford University Press. Barnes, John A. 1994. A Pack of Lies: Towards a Sociology of Lying. Cambridge: Cambridge University Press. Beaver, David and Henk Zeevat. 2007. Accommodation. In Linguistic Interfaces, Gillian Ramchand and Charles Reiss, (eds.), 503–538. Oxford: Oxford University Press. Beyssade, Claire and Jean-Marie Marandin. 2006. The Speech Act Assignment Problem Revisited: Disentangling Speaker’s Commitment from Speaker’s Call on Addressee. In Empirical Issues in Syntax and semantics 6, O. Bonami and P. Cabredo Hofherr (eds.), 37–68. Bianchi, Claudia (ed.). 2004. The Semantics/Pragmatics Distinction. Stanford: CSLI Publications. Birkner, Karin. 2008. Relativ(satz)konstruktionen im gesprochenen Deutsch. Syntaktische, prosodische, semantische und pragmatische Aspekte. Berlin, New York: De Gruyter. Black, Max. 1952. Saying and Disbelieving. Analysis 13: 25–33. Black, Max. 1983. The Prevalence of Humbug. In The Prevalence of Humbug and Other Essays, Max Black, 115–143. Ithaca, London: Cornell University Press. Blome-Tillmann, Michael. 2008. Conversational implicature and the cancellability test. Analysis 68.2: 156–160. Blühdorn, Hardarik. 2008. Zur Struktur und Interpretation von Relativsätzen. Deutsche Sprache 35: 287–314 Blum, Susan D. 2005. Five approaches to explaining “truth” and “deception” in human communication. Journal of Anthropological Research 61 (3): 289–315. Blum, Susan D. 2007. Lies That Bind. Chinese Truth, Other Truths. Lanham, Maryland: Rowman & Littlefield Publishers. Boisvert, Daniel and Kirk Ludwig. 2006. Semantics for nondeclaratives. In The Oxford Handbook of Philosophy of Language, Ernie Lepore, and Barry Smith (eds.), 864–892. Oxford: Oxford University Press. Bolinger, Dwight. 1973. Truth is a linguistic question. Language 49: 539–550. Bok, Sissela. 1999. Lying. Moral Choice in Public and Private Life. New York: Random House. Borg, Emma. 2004. Minimal Semantics. Oxford: Clarendon Press. Borg, Emma. 2010. Meaning and Context: A Survey of a Contemporary Debate. In The Later Wittgenstein on Language, Daniel Whiting (ed.), 96–113. Basingstoke: Palgrave Macmillan. Borg, Emma. 2012. Pursuing Meaning. Oxford: Oxford University Press.

240   

   References

Bosco, Francesca M. and Monica Bucciarelli. 2008. Simple and complex deceits and ironies. Journal of Pragmatics 40: 583–607. Brandt, Margareta, Marga Reis, Inger Rosengren und Ilse Zimmermann. 1992. Satztyp, Satzmodus und Illokution. In Satz und Illokution, Bd. 1, Rosengren, Inger (ed.), 1–90. Tübingen: Niemeyer. Brandner, Ellen. 2010. The syntax of verb-initial exclamatives. Studia Linguistica 64: 81–115. Brandom, Robert. 1994. Making it explicit: reasoning, representing, and discursive commitment. Cambridge, Mass.: Harvard University Press. Brendel, Elke, Jörg Meibauer, and Markus Steinbach (eds.). 2011. Exploring the Meaning of Quotation. In Understanding Quotation. Linguistic and Philosophical Analyses, Elke Brendel, Jörg Meibauer, and Markus Steinbach (eds.), 1–34. Berlin, New York: De Gruyter Mouton. Brown, Jessica and Herman Cappelen. 2011. Assertion: An Introduction and Overview. In Assertion. New Philosophical Essays, Jessica Brown and Herman Cappelen (eds.), 1–17. Oxford: Oxford University Press. Brown, Penelope. 2002. Everyone has to lie in Tzeltal. In Talking to Adults: the Contribution of Multiparty Discourse to Language Acquisition, Shoshana Blum-Kulka and Catherine E. Snow, (eds.), 241–275. Mahwa, NJ: Lawrence Erlbaum. Brown, Penelope and Stephen Levinson. 1987. Politeness: Some Universals in Language Usage. Cambridge: Cambridge University Press. Burton-Roberts, Noël. 2010. Cancellation and Intention. In Explicit Communication. Robyn Carston’s Pragmatics, Belén Soria and Esther Romero (eds.), 138–155. Basingstoke: Palgrave Macmillan. Burton-Roberts, Noël. 2013. On Grice and cancellation. Journal of Pragmatics 48: 17–28. Capone, Alessandro. 2009. Are Explicatures Cancellable? Toward a Theory of the Speaker’s Intentionality. Intercultural Pragmatics 6 (1): 55–83. Cappelen, Herman. 2011. Against Assertion. In Assertion. New Philosophical Essays, Jessica Brown and Herman Cappelen (eds.), 21–48. Oxford: Oxford University Press. Cappelen, Herman and Ernie Lepore. 2004. Insensitive Semantics. Oxford: Blackwell. Carrión, Ricardo E., Julian P. Keenan, and Natalie Sebanz. 2010. A truth that’s told with bad intent: An ERP study of deception. Cognition 114: 105–110. Carson, Thomas L. 2010. Lying and Deception: Theory and Practice. Oxford: Oxford University Press. Carston, Robyn. 2002. Thoughts and Utterances. The Pragmatics of Explicit Communication. Oxford: Blackwell. Carston, Robyn. 2010. Explicit Communication and ‘Free’ Pragmatic Enrichment. In Explicit Communication. Robyn Carston’s Pragmatics, Belén Soria and Esther Romero (eds.), 217–286. Basingstoke: Palgrave Macmillan. Castelfranchi, Cristiano and Isabella Poggi. 1994. Lying as pretending to give information. In Pretending to communicate, Herman Parret (ed.), 276–291. Berlin, New York: De Gruyter. Castelfranchi, Cristiano and Isabella Poggi. 1998. Bugie, finzioni, sotterfugi. Per una scienza dell’inganno. Roma: Carocci. Channell, Joanna. 1994. Vague Language. Oxford: Oxford University Press. Chierchia, Gennaro and Sally McConnell-Ginet. 1990. Meaning and Grammar. An Introduction to Semantics. Cambridge, Mass.: MIT Press. Chisholm, Roderick M. 1976. Person and Object. La Salle, Ill.: Open Court. Chisholm, Roderick M. 1989. Theory of Knowledge. 3rd ed. Englewood Cliffs, NJ: Prentice-Hall.

References   

   241

Chisholm, Roderick M. and Thomas D. Feehan. 1977. The intent to deceive. The Journal of Philosophy 74: 143–159. Claridge, Claudia. 2011. Hyperbole in English. A Corpus-based Study of Exaggeration. Cambridge: Cambridge University Press. Clark, Herbert H. and Richard J. Gerrig. 1990. Quotations as demonstrations. Language 66: 764–805. Cliffe, Lionel, Maureen Ramsay, and Dave Bartlett. 2000. The Politics of Lying. Implications for Democracy. Basingstoke: Macmillan. Cohen, G.A. 2002. Deeper into Bullshit. In The Contours of Agency: Essays on Themes from Harry Frankfurt, Sarah Buss and Lee Overton (eds.), 321–339. Cambridge, Mass.: MIT Press. Cohen, Ted. 1973. Illocutions and perlocutions. Foundations of Language 9: 492–503. Cole, Shirley A.N. 1996. Semantic prototypes and the pragmatics of lie across cultures. The LACUS Forum 23: 475–483. Coleman, Linda. 2007. True Lies: Collusion, Evasion, and the Assignment of Responsibility in Cases of Misrepresentation. In Cultures of Lying. Theories and Practice of Lying in Society, Literature, and Film, Jochen Mecke (ed.), 47–68. Berlin: Galda + Wilch. Coleman, Linda and Paul Kay. 1981. Prototype Semantics: The English Verb lie. Language 57: 26–44. Colston, Herbert L. and Shauna B. Keller. 1998. You’ll never believe this: Irony and hyperbole in expressing surprise. Journal of Psycholinguistic Research 27: 499–513. Conrad, Sarah-Jane. 2011. Disquotational Indirect Reports in Focus. In Understanding Quotation, Elke Brendel, Jörg Meibauer, and Markus Steinbach (eds.), 59–84. Berlin, New York: De Gruyter Mouton. Cresswell, Max. 1982. The autonomy of semantics. In Processes, Beliefs, and Questions, Stanley Peters and Esa Saarinen (eds.), 69–86. Dordrecht: Kluwer. Creusere, Marlena A. 2000. A Developmental Test of Theoretical Perspectives on the Understanding of Verbal Irony: Children’s Recognition of Allusion and Pragmatic Insincerity. Metaphor and Symbol 15: 29–45. Cummings, Louise. 2009. (ed.) The Pragmatics Encyclopedia. London: Routledge. Curme, George. 1922. A Grammar of the German Language. New York: Frederick Ungar Publ. Currie, Gregory. 1990. The Nature of Fiction. Cambridge: Cambridge University Press. Danziger, Eve. 2010. On trying and lying: Cultural configurations of Grice’s Maxim of Quality. Intercultural Pragmatics 7 (2): 199–219. Davidson, Donald. 1984. Inquiries into truth and interpretation. Oxford: Oxford University Press. Davidson, David. 1986. Deception and Division. In The Multiple Self, Jon Elster (ed.), 79–92. Cambridge: Cambridge University Press. d’Avis, Franz. 2013. Exklamativsatz. In Satztypen des Deutschen, Jörg Meibauer, Markus Steinbach, und Hans Altmann (eds.), 171–201. Berlin, New York: De Gruyter. Davis, Wayne A. 1998. Implicature. Intention, convention, and principle in the failure of Gricean theory. Cambridge: Cambridge University Press. Davis, Wayne A. 2007. How normative is implicature? Journal of Pragmatics 39: 1655–1672. Deemter, Kees van. 2010. Not exactly: In Praise of Vagueness. Oxford: Oxford University Press. DePaulo, Bella M., J.J. Lindsay, B.E. Malone, K. Muhlenbruck, K. Charlton, and H. Cooper. 2003. Cues to deception. Psychological Bulletin 129: 74–118.

242   

   References

Deweese-Boyd, Ian. 2006. Self-Deception. Stanford Encyclopedia of Philosophy. [http://plato. stanford.edu/ entries/self-deception] Dietz, Simone. 2002. Der Wert der Lüge: über das Verhältnis von Sprache und Moral. Paderborn: Mentis. du Boulay, Juliet. 1976. Lies, mockery and family integrity. In Mediterranean family structures, Peristiany, Jean George (ed.), 389–406. Cambridge: Cambridge University Press. Doran, Ryan, Gregory Ward, Meredith Larson, Yaron McNabb, and Rachel E. Baker. 2012. A novel experimental paradigm for distinguishing between what is said and what is implicated. Language 88: 124–154. Dov Lerner, Berel. 2009. Review of Colin McGinn, Mindfucking: A Critique of Mental Manipulation. Philosophy in Review 29 (2): 123–124. Dynel, Marta. 2011. A web of deceit: A neo-Gricean view on types of verbal deception. International Review of Pragmatics 3 (2): 137–165. Evans, Mark. 2006. The Republic of Bullshit: On the Dumbing-Up of Democracy. In Bullshit and Philosophy. Guaranteed to Get Perfect Results Every Time. Gary L. Hardcastle and George A. Reisch (eds.), 185–202. Chicago/La Salle, Illinois: Open Court. Falkenberg, Gabriel. 1982. Lügen. Grundzüge einer Theorie sprachlicher Täuschung. Tübingen: Niemeyer. Falkenberg, Gabriel. 1990. Searle on sincerity. In Speech Acts, Meaning, and Intentions. Critical Approaches to the Philosophy of John R. Searle, Armin Burkhardt (ed.), 129–146. Berlin, New York: De Gruyter. Faller, Martina. 2012. Evidential scalar implicatures. Linguistics and Philosophy 35 (4): 285–312. Fallis, Don. 2009. What is lying? The Journal of Philosophy 106 (1): 29–56. Fallis, Don. 2011. Lying and Deception. Philosopher’s Imprint 10: 1–22 Fallis, Don. 2012. Lying as a Violation of Grice’s Maxim of Quality. Dialectica 66 (4): 563–581. Fallis, Don. 2013. Are Bald-Faced Lies Deceptive After All? (April 12, 2013). Available at SSRN (Social Science Research Network: http://ssrn.com/abstract=2250050 or http://dx.doi. org/10.2139/ssrn.2250050 Fónagy, Iván. 2001. Languages within Language: An Evolutive Approach. Amsterdam, Philadelphia: Benjamins. Forbes, Graeme. 1997. How much substitutivity? Analysis 57: 109–113. Forbes, Graeme. 1999. Enlightened semantics for simple sentences. Analysis 59: 86–91. Frankfurt, Harry G. 2002. Reply to Cohen. In The Contours of Agency: Essays on Themes from Harry Frankfurt, Sarah Buss and Lee Overton (eds.), 340–344. Cambridge, Mass.: MIT Press. Frankfurt, Harry G. 2005. On Bullshit. Princeton, Oxford: Princeton University Press. Fraser, Bruce. 1988. Motor oil is motor oil. An account of English nominal tautologies. Journal of Pragmatics 12: 215–220. Freienstein, Jan Claas. 2008. Das erweiterte Appositiv. Tübingen: Narr. Fricker, Elizabeth. 2012. Stating and Insinuating. Proceedings of the Aristotelian Society Supplementary Volume LXXXVI: 61–94. Garrett, Merrill and Robert Harnish. 2007. Experimental pragmatics: testing for implicitures. Pragmatics & Cognition 15: 65–90. Gazdar, Gerald. 1981. Speech act assignment. In Elements of Discourse Understanding, Aravind Joshie, Bonnie Webber, and Ivan A. Sag (eds.), 64–83. Cambridge: Cambridge University Press.

References   

   243

Geurts, Bart. 2010. Quantity Implicatures. Oxford: Oxford University Press. Gilsenan, Michael. 1976. Lying, honor and contradiction. In Transaction and meaning: directions in the anthropology of exchange and symbolic behaviour, Bruce Kapferer (ed.), 191–219. Philadelphia: Institute for the Study of Human Issues. Grewendorf, Günther. 1982. Behaupten und Zustimmen. In Logik und Pragmatik. Zum Rechtfertigungsproblem logischer Sprachregeln, Carl Friedrich Gethmann, (ed.), 125–142. Frankfurt/M.: Suhrkamp. Grice, Paul. 1989a. Logic and Conversation. In: Studies in the Way of Words, Paul Grice. 22–40. Cambridge, Mass.: Harvard University Press. Grice, Paul. 1989b. Further Notes on Logic and Conversation. In Studies in the Way of Words, Paul Grice. 41–57. Cambridge, Mass.: Harvard University Press. Grice, Paul. 1989c. Meaning. In Studies in the Way of Words, Paul Grice. 213–223. Cambridge, Mass.: Harvard University Press. Grice, Paul. 1989d. Retrospective Epilogue. In Studies in the Way of Words, Paul Grice. 339–385. Cambridge, Mass.: Harvard University Press. Grimshaw, Elizabeth. 1979. Complement selection and the lexicon. Linguistic Inquiry 10: 279–326. Gross, Steven. 2001. Essays on Linguistic Context-Sensitivity and its Philosophical Significance. New York, London: Routledge. Grosz, Patrick. 2012. On the Grammar of Optative Constructions. Amsterdam, Philadelphia: Benjamins. Grundy, Peter. 2000. Doing Pragmatics. London: Arnold. Gunlogson, Christine. 2003. True to Form: Rising and Falling Declaratives as Questions in English. Ph.D. Diss., University of California, Santa Cruz. Gupta, Swati, Kayo Sakamoto, and Andrew Ortony. 2012. Telling it like it isn’t – a comprehensive approach to analyzing deception. In The goals of cognition. Festschrift for Cristiano Castelfranchi, Fabio Paglieri, Luca Tummolini, Rino Falcone, and Maria Miceli (eds.), 579–610. London: College Publications. Gutzmann, Daniel and Erik Stei. 2011. Quotation Marks and Kinds of Meaning. Arguments in Favor of a Pragmatic Account. In Understanding Quotation, Elke Brendel, Jörg Meibauer, and Markus Steinbach (eds.), 161–194. Berlin, New York: De Gruyter Mouton. Hancock, J.T., J. Thom-Santelli and T. Ritchie. 2004. Deception and design: The impact of computer technologies in lying behaviour. In Proceedings of the Conference on Computer Human Communication, 130–136. New York: ACM Press. Harada, Tokiko, Shoji Itakura, Fen Xu, Kang Lee, Satoru Nakashita, Daisuke N. Saito, and Norihiro Sadato. 2009. Neural correlates of the judgment of lying: A functional magnetic resonance imaging study. Neuroscience Research 63 (1): 24–34. Hardin, Karol J. 2010. The Spanish notion of Lie: Revisiting Coleman and Kay. Journal of Pragmatics 42: 3199–3213. Harnish, Robert M. 1990. Speech Acts and Intentionality. In Speech Acts, Meaning, and Intention, Armin Burkhardt (ed.), 169–193. Berlin, New York: de Gruyter. Harnish, Robert M. 1994. Mood, meaning and speech acts. In Foundations of Speech Act Theory, Savas L. Tsohatzidis (ed.), 407–459. London, New York: Routledge. Harras, Gisela, Edeltraud Winkler, Sabine Erb und Kristel Proost. 2004. Handbuch deutscher Kommunikationsverben. Teil I: Wörterbuch. Berlin, New York: de Gruyter. Harweg, Roland. 1972. Reduzierte Rede. Linguistics 80: 44–55.

244   

   References

Henst, Jean-Baptiste van der, Laure Carles, and Dan Sperber. 2002. Truthfulness and Relevance in Telling the Time. Mind and Language 17: 457–466. Hirstein, William. 2005. Brain Fiction: Self-Deception and the Riddle of Confabulation. Cambridge, Mass.: MIT Press. Höhle, Tilman N. 1992. Über Verum-Fokus im Deutschen. In Informationsstruktur und Grammatik. Linguistische Berichte, Sonderheft 4/1991–1992, Joachim Jacobs (ed.), 112–141. Opladen: Westdeutscher Verlag. Hoffmann, Ludger. 2008. Über ja. Deutsche Sprache 36: 193–219. Holler, Anke. 2007. Uniform oder different? Zum syntaktischen Status nicht-restriktiver Relativsätze. Deutsche Sprache 35: 250–270. Horn, Laurence R. 2004. Implicature. In The Handbook of Pragmatics, Laurence R. Horn and Gregory Ward (eds.), 3–28. Oxford: Blackwell. Horn, Laurence R. 2008. On F-implicature. Myth-analysis and rehabilitation. Ms. New Haven, CT. Horn, Laurence R. 2009. WJ-40: Implicature, truth, and meaning. International Review of Pragmatics 1: 3–34. Huang, Yan. 2012. The Oxford Dictionary of Pragmatics. Oxford: Oxford University Press. Jary, Mark. 2010. Assertion. Basingstoke: Palgrave Macmillan. Jaszczolt, Katarzyna M. 2005. Default Semantics: Foundations of Compositional Theory of Acts of Communication. Oxford: Oxford University Press. Jaszczolt, Katarzyna M. 2009. Cancellability and the primary/secondary meaning distinction. Intercultural Pragmatics 6 (3): 259–289. Johnson, Michael and Ernie Lepore. 2011. Misrepresenting misrepresentation. In Understanding Quotation, Elke Brendel, Jörg Meibauer, and Markus Steinbach (eds.), 231–248. Berlin, Boston: De Gruyter Mouton. Kant, Immanuel. 1968. Über ein vermeintes Recht aus Menschenliebe zu lügen. In Immanuel Kant. Werke in zwölf Bänden, VIII: Schriften zur Ethik und Religionsphilosophie 2, Wilhelm Weischedel (ed.), 637–643. Frankfurt/M.: Suhrkamp. Kaplan, David. 1999. The Meaning of ouch and oups. Explorations in the theory of Meaning as Use. Unpublished Ms. University of California, Los Angeles, USA. Katsos, Napoleon. 2012. Experimental investigations and pragmatic theorising. In The Cambridge Handbook of Pragmatics, Keith Allan and Kasia M. Jaszczolt (eds.), 275–290. Cambridge: Cambridge University Press. Kenyon, Tim. 2003. Cynical assertion: Convention, Pragmatics, and Saying ‘Uncle’. American Philosophical Quarterly 40 (3): 241–248. Kissine, Mikhail. 2008. Assertoric commitment. In Commitment (Belgian Journal of Linguistics 22), Philippe D. Brabanter and Patrick Dendale (eds.), 155–177. Kissine, Mikhail. 2012. Sentences, utterances, and speech acts. In The Cambridge Handbook of Pragmatics, Keith Allan and Kasia M. Jaszczolt (eds.), 169–190. Cambridge: Cambridge University Press. Kissine, Mikhail. 2013. From utterances to speech acts. Oxford: Oxford University Press. Klockow, Reinhold. 1980. Linguistik der Gänsefüßchen. Untersuchungen zum Gebrauch der Anführungszeichen im gegenwärtigen Deutsch. Frankfurt/M.: Haag und Herchen. König, Ekkehard and Peter Siemund. 2007. Speech act distinctions in grammar. In: Language Typology and Syntactic Description, Shopen, Timothy (ed.), 276–324. Cambridge: Cambridge University Press.

References   

   245

Kratzer, Angelika. 1999. Beyond Ouch and Oops. How descriptive and expressive meaning interact. Unpublished Ms. Krifka, Manfred. 2002. Be Brief and Vague! And how Bidirectional Optimality Theory allows for Verbosity and Precision. In Sounds and Systems. Studies in Structure and Change. A Festschrift for Theo Vennemann, David Restle and Dietmar Zaefferer (eds.), 439–458. Berlin, New York: De Gruyter. Krifka, Manfred. 2011. Questions. In Semantics. An International Handbook of Natural Language Meaning, Klaus von Heusinger, Claudia Maienborn, and Paul Portner (eds.) (2011), 1742–1785. Berlin, Boston: De Gruyter Mouton. Kümmerling-Meibauer, Bettina and Jörg Meibauer. 2011. Lügenerwerb und Geschichten vom Lügen. LiLi. Zeitschrift für Literaturwissenschaft und Linguistik 162: 118–138. Künne, Wolfgang. 2003. Conceptions of Truth. Oxford: Oxford University Press. Lackey, Jennifer. 2008. Learning from Words. Testimony as a Source of Knowledge. Oxford: Oxford University Press. Lackey, Jennifer. 2013. Lies and deception: an unhappy divorce. Analysis 73 (2): 236–248. Lang, Ewald. 1982. Einstellungsausdrücke und ausgedrückte Einstellungen. In Untersuchungen zur Semantik. (studia grammatica XXII), Rudolph Růžička und Wolfgang Motsch (Hgg.), 305–341. Berlin: Akademie-Verlag. Lapp, Edgar. 1992. Linguistik der Ironie. Tübingen: Narr. Lasersohn, Peter. 1999. Pragmatic Halos. Language 75: 522–551. Lasersohn, Peter. 2012. Contextualism and compositionality. Linguistics and Philosophy 35: 171–189. Lee, Kang. 2013. Little Liars: Development of Verbal Deception in Children. Child Development Perspectives 7 (2): 91–96. Lee, Kang and Victoria Talwar. 2014. Children and Lying: A Century of Scientific Research. Hoboken, NJ: Wiley-Blackwell. Leekam, Susan. 1991. Jokes and Lies: Children’s Understanding of Intentional Falsehood. In Natural Theories of Mind. Evolution, Development amd Simulation of Everyday Mindreading, Andrew Whiten (ed.), 159–174. Oxford: Blackwell. Letnes, Ole. 2002. Wollen: Zwischen Referat und Verstellung. In Modus, Modalverben, Modalpartikeln. Catherine Fabricius-Hansen, Oddleif Leirbukt und Ole Letnes (eds.), 101–121. Trier: WVT. Levinson, Stephen C. 1983. Pragmatics. Cambridge: Cambridge University Press. Levinson, Stephen C. 2000. Presumptive Meanings. The Theory of Generalized Conversational Implicature. Cambridge, Mass.: MIT Press. Levine, Emma E. and Maurice E. Schweitzer (2013): Prosocial Lies: When Deception Breeds Trust. Ms. Philadelphia: University of Pennsylvania [http://ssrn.com/abstract= 2266091] Lewis, David. 1979. Scorekeeping in a Language game. Journal of Philosophical Logic 8: 339–359. Lindner, Katrin. 1991. Wir sind ja doch alte Bekannte. The use of ja and doch as modal particles. In Discourse Particles, Abraham, Werner (ed.), 163–201. Amsterdam, Philadelphia: Benjamins, 163–201. Lohnstein, Horst. 2012. Verumfokus – Satzmodus – Wahrheit. In Wahrheit – Fokus – Negation, Hardarik Blühdorn and Horst Lohnstein (eds.), 31–67. Hamburg: Buske. Loock, Rudy. 2007. Appositive relative clauses and their functions in discourse. Journal of Pragmatics 39: 336–362. MacFarlane, John. 2007. Relativism and Disagreement. Philosophical Studies 132: 17–31

246   

   References

MacFarlane, John. 2009. Nonindexical Contextualism. Synthese 166, 231–250. MacFarlane, John. 2011. What Is Assertion? In Assertion. New Philosophical Essays, Jessica Brown and, Herman Cappelen (eds.), 79–96. Oxford: Oxford University Press. Maes, Hans and Katrien Schaubroeck. 2006. Different Kinds and Aspects of Bullshit. In Bullshit and Philosophy. Guaranteed to Get Perfect Results Every Time, Gary L. Hardcastle and George A. Reisch (eds.), 171–183. Chicago/La Salle, Illinois: Open Court. Mahon, James Edwin. 2008. The definition of lying and deception. Stanford Encyclopedia of Philosophy. Marcu, Daniel. 2000. Perlocutions: The Achilles’ heel of speech act theory. Journal of Pragmatics 32: 1719–1741. Mateo, José and Francisco Yus. 2013. Towards a cross-cultural pragmatic taxonomy of insults. Journal of Language Aggression and Conflict 1 (1): 87–114. Mayol, Laia and Elena Castroviejo. 2013. How to cancel an implicature. Journal of Pragmatics 50: 84–104. McGinn, Colin. 2008. Mindfucking. A Critique of Mental Manipulation. Stocksfield: Acumen. McGlone, Matthew. 2009. Deception by selective quotation. In The interplay of truth and deception. New agendas in communication, Matthew McGlone and Mark L. Knapp (eds.), 54–65. London: Routledge. McCready, Eric. 2010. Varieties of conventional implicature. Semantics & Pragmatics 3: 1–57. McCready, Eric. 2012. Emotive equilibria. Linguistics and Philosophy 35: 243–283. Mecke, Jochen. 2007. Cultures of Lying. In Cultures of Lying. Theories and Practice of Lying in Society, Literature, and Film, Jochen Mecke (ed.), 1–30. Glienicke/Berlin, Madison, Wisconsin: Galda+Wilch. Meibauer, Jörg. 1986. Rhetorische Fragen. Tübingen: Niemeyer. Meibauer, Jörg. 1991. Existenzimplikaturen bei rhetorischen W-Fragen. In Fragen und Fragesätze, Marga Reis und Inger Rosengren, Inger (eds.), 223–242. Tübingen: Niemeyer. Meibauer, Jörg. 1994. Modaler Kontrast und konzeptuelle Verschiebung. Studien zur Syntax und Semantik von Modalpartikeln. Tübingen: Niemeyer. Meibauer, Jörg. 1997. Modulare Pragmatik und die Maximen der Modalität. In Pragmatik. Implikaturen und Sprechakte (Linguistische Berichte, Sonderheft 8), Eckard Rolf (ed.), 226–256. Opladen: Westdeutscher Verlag. Meibauer, Jörg. 2005. Lying and falsely implicating. Journal of Pragmatics 37: 1373–1399 Meibauer, Jörg. 2006. Implicature. In Encyclopedia of Language and Linguistics. 2nd edition. Vol. V., Keith Brown (ed.), 568–580. Oxford: Elsevier. Meibauer, Jörg. 2007. Zitat und Lüge. In Zitat und Bedeutung. Linguistische Berichte, Sonderheft 15, Elke Brendel, Jörg Meibauer und Markus Steinbach (eds.), 181–200. Hamburg: Buske. Meibauer, Jörg. 2008. Tautology as Presumptive Meaning. Pragmatics & Cognition 16 (3): 437–468. Meibauer, Jörg. 2011. On lying: intentionality, implicature and imprecision. Intercultural Pragmatics 8 (2): 277–292. Meibauer, Jörg. 2012a. Pragmatic evidence, context, and story design. An essay on recent developments in experimental pragmatics. Language Sciences 34 (6): 768–776. Meibauer, Jörg. 2012b. What is a context? Theoretical and empirical evidence. In What is a Context? Linguistic approaches and challenges, Rita Finkbeiner, Jörg Meibauer, and Petra B. Schumacher (eds.), 9–32. Amsterdam, Philadelphia: Benjamins.

References   

   247

Meibauer, Jörg. 2013a. Satztypen und Pragmatik. In Satztypen des Deutschen, Jörg Meibauer, Markus Steinbach und Hans Altmann (eds.), 711–736. Berlin, Boston: de Gruyter. Meibauer, Jörg. 2013b. Bullshit als pragmatische Kategorie. Linguistische Berichte 235: 267–292. Meibauer, Jörg. 2013c. Bald-faced lies as acts of language aggression. To appear in: Journal of Language Aggression and Conflict Meibauer, Jörg. 2013d. A truth that’s told with bad intent. Lying and implicit content. To appear in: Implicit Content (Belgian Journal of Linguistics 28), Daniela Rossi and Nicolas Ruytenbeek (eds.). Amsterdam, Philadelphia: Benjamins. Meibauer, Jörg, Markus Steinbach and Hans Altmann (eds.). 2013. Satztypen des Deutschen. Berlin, Boston: De Gruyter. Mele, Alfred. 2001. Self Deception Unmasked. Princeton: Princeton University Press. Mortelmans, Tanja and Jeroen Vanderbiesen. 2011. Dies will ein Parlamentarier „aus zuverlässiger Quelle“ erfahren haben. Reportives wollen zwischen sollen und dem Konjunktiv I der indirekten Rede. In Modalität und Evidentialität – Modality and Evidentiality, Gabriele Diewald und Elena Smirnova (ed.), 69–88. Trier: WVT Wissenschaftlicher Verlag Trier. Mooney, Annabelle. 2004. Co-operation, violation and making sense. Journal of Pragmatics 36: 899–920. Moore, Joseph G. 1999. Saving substitutivity in simple sentences. Analysis 57: 102–108. Moore, Joseph G. 2000. Did Clinton lie? Analysis 60 (267): 250–254. Narvaez, Darcia. 2010. Moral complexity: The fatal attraction of truthiness and the importance of mature moral functioning. Perspectives on Psychological Science 5 (2): 163–181. Nemesi, Attila L. 2010. Data-gathering methods in research on hyperbole production and interpretation. In The Role of Data at the Semantics-Pragmatics Interface, Enikő Németh T. and Károly Bibok (eds.), 381–417. Berlin, New York: De Gruyter Mouton. Neu, Jerome. 2008. Sticks and Stones. The Philosophy of Insults. Oxford: Oxford University Press. Noveck, Ira A. 2001. When children are more logical than adults: experimental investigations of scalar implicature. Cognition 78: 165–188. Noveck, Ira and Dan Sperber. 2007. The why and how of experimental pragmatics: the case of ‘scalar inferences’. In Pragmatics, Noël Burton-Roberts (ed.), 184–212. Basingstoke, UK: Palgrave Macmillan. Önnerfors, Olaf. 1997. Verb-erst-Deklarativsätze. Grammatik und Pragmatik. Stockholm: Almqvist & Wiksell. Oppenrieder, Wilhelm. 2013. Deklarativsatz. In Satztypen des Deutschen, Jörg Meibauer, Markus Steinbach und Hans Altmann (eds.), 20–50. Berlin, New York: De Gruyter Mouton. Overstreet, Maryann. 2011. Vagueness and hedging. In Pragmatics of Society, Gisle Andersen and Karin Aijmer (eds.), 293–320. Berlin, New York: De Gruyter Mouton. Pagin, Peter. 2007 Assertion. Stanford Encyclopedia of Philosophy. [http://plato.stanford.edu/ entries/ assertion] Partee, Barbara. 1973. The Syntax and Semantics of Quotations. In A Festschrift for Morris Halle, Stephen R. Anderson and Paul Kiparsky (eds.), 410–418. New York: Holt, Rinehart and Winston. Pascal, Fania. 1984. Wittgenstein: A Personal Memoir. In Recollections of Wittgenstein, Rush Rhees (ed.), 12–49. Oxford: Oxford University Press.

248   

   References

Poschmann, Claudia. 2009. Echo-Fragen. Eine fokusbasierte Metarepräsentationsanalyse. Phil. Diss., Johannes Gutenberg University Mainz. Posner, Roland. 1980. Semantics and pragmatics of sentence connectives in natural language. In Speech Act Theory and Pragmatics, John R. Searle, Ferenc Kiefer, and Manfred Bierwisch (eds.), 168–203. Dordrecht: Reidel. Potts, Christopher. 2005. The logic of conventional implicature. Oxford: Oxford University Press. Potts, Christopher. 2007. The expressive dimension. Theoretical linguistics 33: 165–197. Potts, Christopher. 2012. Conventional implicature and expressive content. In Semantics: An International Handbook of Natural Language Meaning, Vol. 3, Claudia Maienborn, Klaus von Heusinger, and Paul Portner (eds.), 2516–2536. Berlin, New York: De Gruyter Mouton. Predelli, Stefano. 2003a. Scare quotes and their relation to other semantic issues. Linguistics and Philosophy 26: 1–28. Predelli, Stefano. 2003b. ‘Subliminable’ Messages, Scare Quotes, and the Use Hypothesis. In Hybrid Quotations (Belgian Journal of Linguistics 17), Philippe de Brabanter (ed.), 153–166. Amsterdam, Philadelphia: Benjamins. Predelli, Stefano. 2010. From the Expressive to the Derogatory: On the Semantic Role for Non-Truth-Conditional Meaning. In New Waves in Philosophy of Language, Sarah Sawyer (ed.), 164–185. Basingstoke: Palgrave Macmillan. Reboul, Anne. 1994. The description of lies in speech act theory. In Pretending to Communicate, Herman Parret (ed.), 292–298. Berlin, New York: De Gruyter. Recanati, François. 2000. Oratio Obliqua, Oratio Recta. An Essay on Metarepresentation. Cambridge, Mass.: MIT Press. Recanati, François. 2001. Open Quotation. Mind 110: 637–687. Recanati, François. 2004. Literal Meaning. Cambridge: Cambridge University Press. Recanati, François. 2010. Truth-conditional pragmatics. Oxford: Oxford University Press. Reimer, Marga. 1995. Performative Utterances. A Reply to Bach and Harnish. Linguistics and Philosophy 18: 655–675 Reimer, Marga. 2003. Too Counter-Intuitive to Believe? Pragmatic Accounts of Mixed Quotation. In Hybrid Quotations (Belgian Journal of Linguistics 17), Philippe de Brabanter (ed.), 167–186. Amsterdam, Philadelphia: Benjamins. Reis, Marga. 1991. Echo-w-Sätze und Echo-w-Fragen. In Fragesätze und Fragen, Marga Reis und Inger Rosengren (eds.), 49–76. Tübingen: Niemeyer. Reis, Marga. 1999. On Sentence Types in German. An Enquiry into the Relationship between Grammar and Pragmatics. Interdisciplinary Journal for Germanic Linguistics and Semiotic Analysis 4 (2): 195–236. Reis, Marga. 2000. Anmerkungen zu Verb-erst-Satz-Typen im Deutschen. In Deutsche Grammatik in Theorie und Praxis, Rolf Thieroff et al. (eds.), 215–227. Tübingen: Niemeyer. Reis, Marga. 2012. On the Analysis of Echo Questions. Tampa Papers in Linguistics 3: 1–24. Rizzi, Luigi. 1997. The fine structure of the left periphery. In Elements of Grammar. Handbook in Generative Syntax, Liliane Haegemann (ed.), 281–337. Dordrecht: Kluwer. Rhodes, Russell. 2009. A Cross-linguistic Comparison of Tautological Constructions with Special Focus on English. Unpublished Ms. University of California, Berkeley, USA. Roguska, Magdalena. 2007a. Exklamation und Negation. Berlin: Logos. Roguska, Magdalena. 2007b. Echo and Mention in the Dan Sperber and Deirdre Wilson’s Theory of Irony – Polemics. Linguistische Berichte 211: 309–330. Rolf, Eckard. 1997. Illokutionäre Kräfte. Grundbegriffe der Illokutionslogik. Opladen: Westdeutscher Verlag.

References   

   249

Roncador, Manfred von. 1988. Zwischen direkter und indirekter Rede. Nichtwörtliche direkte Rede, erlebte Rede, logophorische Konstruktionen und Verwandtes. Tübingen: Niemeyer. Rott, Hans. 2003. Der Wert der Wahrheit. In Kulturen der Lüge, Mathias Mayer (ed.), 7–34. Köln, Weimar: Böhlau. Saarni, Carolyn and Michael Lewis. 1993. Deceit and Illusion in Human Affairs. In Lying and Deception in Everyday Life, Michael Lewis and Carolyn Saarni (eds.), 1–29. New York, London: Guilford Press. Sacks, Harvey. 1975. Everyone has to lie. In Sociocultural dimensions of language use, M. Sanches and B.G. Blount (eds.), 57–80. New York: Academic Press. Sainsbury, Richard Mark. 2010. Fiction and Fictionalism. London, New York: Routledge. Saka, Paul. 1998. Quotation and the Use-Mention Distinction. In: Mind 107: 114–136. Saka, Paul. 2003. Quotational Constructions. In Hybrid Quotations (Belgian Journal of Linguistics 17). Philippe de Brabanter (ed.), 187–212. Amsterdam, Philadelphia: Benjamins. Saul, Jennifer. 2000. Did Clinton say something false? Analysis 60 (267): 255–257. Saul, Jennifer. 2002. Speaker meaning, what is said, and what is implicated. Noûs 36: 228–248. Saul, Jennifer. 2012. Lying, Misleading, and What is Said. Oxford: Oxford University Press. Searle, John R. 1969. Speech Acts. An Essay in the Philosophy of Language. Cambridge: Cambridge University Press. Searle, John R. 1979a. Indirect speech acts. In Expression and Meaning. Studies in the Theory of Speech Acts, John R. Searle, 30–57. Cambridge: Cambridge University Press. Searle, John R. 1979b. The logical status of fictional discourse. In Expression and Meaning. Studies in the Theory of Speech Acts, John R. Searle, 58–75. Cambridge: Cambridge University Press. Searle, John R. 1989. How performatives work. Linguistics and Philosophy 12: 535–558. Schiffer, Stephen. 1972. Meaning. Oxford: Clarendon Press. Schlenker, Philippe. 2007. Expressive presuppositions. Theoretical Linguistics 33 (2): 237–245. Schlesinger, Izchak and Sharon Hurvitz. 2008. The structure of misunderstandings. Pragmatics & Cognition 16 (3): 568–585. Simons, Mandy. 2003. Presupposition and accommodation: Understanding the Stalnakerian picture. Philosophical Studies 112: 251–278. Simons, Mandy. 2005. Presupposition and Relevance. In Semantics versus Pragmatics, Zoltán Gendler Szabó (ed.), 329–355. Oxford: Oxford University Press. Simpson, David. 1992. Lying, Liars and Language. Philosophy and Phenomenological Research LII: 623–639. Sip, Kamila E., Andreas Roepstorff, William McGregor, and Chris D. Firth. 2008. Detecting deception: the scope and limits. Trends in Cognitive Sciences 12: 48–53. Smith, Barry C. 2010. Relativism, disagreement, and predicates of personal taste. In Contextdependence, perspective and reality, François Recanati, Isidora Stojanovic, and Neftalí Villanueva, (eds.), 195–223. Berlin, New York: De Gruyter Mouton. Socka, Anna. 2011. Evidentialität und Epistemizität in der Bedeutung reportativer Satzadverbien im Polnischen und Deutschen. In Modalität und Evidentialität – Modality and Evidentiality, Gabriele Diewald and Elena Smirnova (eds.), 49–68. Trier: WVT. Sorensen, Roy. 2007. Bald-faced lies! Lying without the intent to deceive. Pacific Philosophical Quarterly 88: 251–264. Sorensen, Roy. 2010. Knowledge-Lies. Analysis 70 (4): 608–615.

250   

   References

Sorensen, Roy. 2012. Lying with Conditionals. The Philosophical Quarterly 62: 820–832. Spence, Sean A., Mike D. Hunter, Tom F.D. Farrow, Russell D. Green, David H. Leung, Catherine J. Hughes, and Venkatasubramanian Ganesan. 2004. A cognitive neurobiological account of deception: Evidence from functional neuroimaging. Philosophical Transactions of the Royal Society of London 359: 1755–1762. Sperber, Dan and Deirdre Wilson. 1981. Irony and the Use-Mention-Distinction. In Radical Pragmatics, Peter Cole (ed.), 295–318. New York: Academic Press. Sperber, Dan and Deirdre Wilson. 1995. Relevance: communication and cognition. 2nd edition. Oxford: Blackwell. Sperber, Dan and Deirdre Wilson. 1998. Irony and relevance: A reply to Seto, Hamamoto and Yamanashi. In Relevance theory. Applications and implications, Robyn Carston and Seiji Uchida (eds.), 283–293. Amsterdam, Philadelphia: Benjamins. Stainton, Robert. 1999. Remarks on the syntax and semantics of mixed quotation. In Philosophy and Linguistics, R. Murasugi and Robert Stainton. (eds.), 259–278. Boulder, Colorado: Westview Press. Stainton, Robert. 2006. Words and Thoughts. Subsentences, Ellipsis, and the Philosophy of Language. Oxford: Oxford University Press. Stalnaker, Robert. 1979. Assertion. In Syntax and Semantics 9. Pragmatics, Peter Cole (ed.), 315–332. New York: Academic Press. Stalnaker, Robert. 2002. Common Ground. Linguistics and Philosophy 25: 701–721. Stanley, Jason. 2008. Knowledge and Certainty. Philosophical Issues 18: 35–57. Stokke, Andreas. 2013a. Lying, Deceiving, and Misleading. Philosophy Compass 8: 348–359. Stokke, Andreas. 2013b. Lying and Asserting. Journal of Philosophy CX(1): 33–60 Stokke, Andreas. 2013c. Saying too Little and Saying too Much. Critical Notice of Lying, Misleading, and What is said by Jennifer Saul. Disputatio V (35): 81–91. Sweetser, Eve V. 1987. The definition of lie. In Cultural models in language and thought, Dorothy Holland and Naomi Quinn (eds.), 3–66. Cambridge: Cambridge University Press. Szabó, Zoltan Gendler. 2005. Semantics versus Pragmatics. Oxford: Clarendon Press. Talwar, Victoria and Kang Lee. 2008. Social and Cognitive Correlates of Children’s Lying Behavior. Child Development 79: 866–881. Terkourafi, Marina. 2010. What-is-Said From Different Points of View. Language and Linguistics Compass 4/8, 705–718. Travis, Charles. 1989. The Uses of Sense: Wittgenstein’s Philosophy of Language. Oxford: Oxford University Press. Tsohatzidis, Savas L. 1990. A few untruths about ‘lie’. In Meaning and Prototypes, Savas L. Tsohatzidis (ed.), 438–446. London, New York: Routledge. Turry, John. 2013. The test of truth: An experimental investigation of the norm of assertion. Cognition 129: 279–291. Unger, Peter. 1975. Ignorance. Oxford: Oxford University Press. Vincent Marelli, Jocelyne. 2004. Words in the Way of Truth. Truthfulness, Deception, Lying Across Cultures and Disciplines. Napoli: Edizioni Scientifiche Italiane. Vincent Marelli, Jocelyne. 2006. Truthfulness. In Handbook of Pragmatics, Jan-Ola Östman, Jef Verschueren and Eline Versluys (eds.), 1–72. Amsterdam, Philadelphia: Benjamins. Vincent, Jocelyne and Cristiano Castelfranchi. 1981. On the art of deception: How to lie while saying the truth. In Possibilities and Limitations of Pragmatics. Proceedings of the Conference on Pragmatics, Urbino, July 8–14, Herman Parret, Marina Sbisà, and Jef Verschueren (eds.), 749–777. Amsterdam, Philadelphia: Benjamins.

References   

   251

Vrij, Aldert. 2008. Detecting Lies and Deceit: The Psychology of Lying and the Implications for Professional Practice. New York: John Wiley & Sons. Wade, Elizabeth and Herbert H. Clark. 1993. Reproduction and demonstration in quotations. Journal of Memory and Language 32: 805–819. Wahrig-Burfeind, Renate. 2006. Wahrig. Deutsches Wörterbuch. 8. Aufl. Gütersloh, München: Bertelsmann. Walczyk, Jeffrey K., Karen S. Roper, Eric Seemann, and Angela M. Humphrey. 2003. Cognitive mechanisms underlying lying to questions: response times as a cue to deception. Applied Cognitive Psychology 17: 755–774. Walton, Kendall L. 1990. Mimesis as Make-Believe. On the Foundations of the Representational Arts. Cambridge, Mass.: Harvard University Press. Ward, Gregory L. and Julia Hirschberg. 1990. A pragmatic analysis of tautological utterances. Journal of Pragmatics 15: 507–520. Weiner, Matthew. 2006. Are all conversational implicatures cancellable? Analysis 66: 127–130. Whitty, Monica T. and S.E. Carville. 2008. Would I lie to you? Self-serving lies and other-oriented lies told across different media. Computers in Human Behaviour 24 (4): 1707–1723. Whitty, Monica T., Adam Joinson. 2009. Truth, Lies and Trust on the Internet. London: Routledge. Wierzbicka, Anna. 1987. Boys will be boys: ‘radical pragmatics’ vs. ‘radical semantics’. Language 63: 95–114. Williams, Bernard. 2002. Truth and truthfulness. An essay in genealogy. Princeton: Princeton University Press. Williams, Emma Joanne, Lewis A. Bott, John Patrick, and Michael B. Lewis. 2013. Telling Lies: The Irrepressible Truth? Plos One 8.4 e60713, 1–14. Williamson, Timothy. 1996. Knowing and Asserting. The Philosophical Review 105: 489–523. Wilson, Deirdre. 2006. The pragmatics of verbal irony: Echo or pretence? Lingua 116: 1722–1743. Wilson, Deirdre and Dan Sperber. 1992. On verbal irony. Lingua 87: 53–76. Wilson, Deirdre and Dan Sperber. 2002. Truthfulness and Relevance. Mind 111: 583–632. Wimmer, Heinz and Josef Perner. 1983. Beliefs about beliefs: representation and constraining function of wrong beliefs in young children’s understanding of deception. Cognition 13: 103–128. Winner, Ellen. 1988. The Point of Words. Children’s Understanding of Metaphor and Irony. Cambridge, Mass.: Harvard University Press. Wittgenstein, Ludwig. 1922. Tractatus logico-philosophicus. Logisch-philosophische Abhandlung. London: Kegan Paul, Trench, Trubner. Wolf, Lari and Ariel Cohen. 2011. Clarity as Objectivized Belief. In Vagueness and Language Use, Paul Égré and Nathan Klinedinst (eds.), 165–190. Basingstoke: Palgrave Macmillan. Zanuttini, Rafaella and Paul Portner. 2003. Exclamative Clauses and the Syntax-Semantics Interface. Language 79: 39–81. Zhang, Fenghui. 2008. Conversational constraint of truthfulness on presuppositions. Intercultural Pragmatics 5 (1): 367–388. Zimmermann, Thomas Ede. 2011. Model-theoretic semantics. In Semantics. An International Handbook of Natural Language Meaning. Vol. 1, Claudia Maienborn, Klaus von Heusinger, and Paul Portner (eds.), 762–802. Berlin, New York: De Gruyter Mouton.

Index accommodation 56, 90–91, 139 accuracy 211–214 – perspectival accuracy 212 acquisition of lying 15–17 acting 26 added parameters approach 39 appearance-reality distinction 15 aspect-sensitivity 156–158 assertion 48, 65–71, 74, 79–82, 88–91, 99–101, 102, 175, 183 – certainty norm for assertion 177 – indirect assertion 91–94 assertive question 51–52 assertives 68 assumption of verbatim reproduction (AVR) 187–190 attachment 200–202 attitude 69, 98–99 attributed thought 163 audience implicature 117–119 availability 117 bald-faced lie 107–109 basic principle of truth-conditional semantics 84 belief 30–31, 68 – first-order belief 15 – objectivised belief 222 – second-order belief 15 belief implicature 77–78 belief set 88 bullshit, bullshitting 9, 10, 29, 171–182, 224, 228 – academic bullshit 180–181 – advertising bullshit 173 – evasive bullshit 176 cancellability 119–120, 125, 145 certainty 113, 176–177, 223–225 – degrees of certainty 81, 113, 225 checklist semantics 41–42 clarity 219–223 – belief theory of clarity 220 – justification theory of clarity 221

command 48 commission 30–31 commitment 66, 74, 79–82, 218 common ground 55, 88–91 compositionality 83 compositional semantics 63 computer-mediated communication (CMC) 10–12 confabulation 111–113 confirmative question 48 conjecture 80, 82, 161 conjunction buttressing 128, 131–132 consistency 19 constitutive rule 67–68 context 35, 38–39, 78, 122, 157, 169, 236 – context set 88 – context shifting argument 38–39 – coordination of contexts 214 – warranting context 106, 147 contextualism 38–39 contradiction 19 control lie 24 cooperative presumption condition 114 cooperative principle 15, 34 cynical assertion 108–109 deception, deceiving 8, 25–32, 104–106 declaration 93 declarative sentence 49–50 defeasibility 89–90 deliberate ambiguity 29 demonstration 187–190 describing 187 descriptive ineffability 96 determinacy condition 114 disagreement 211–214 – faultless disagreement 151 – lost disagreement 211 discourse adverb 219–223 doubt 113 echo question 52–53 ellipsis 70–71 essential condition 72

Index   

essential rule 74 ethics 14–15, 154 event related potentials (ERP) 19–21 evidentiality hedges 80–81 exclamation 48 exclamative sentence 54–59 executive functioning 16 experimental pragmatics 17, 236 explicature 37, 140–147 explicit performative 91–94 expressing an attitude 69 expressive meaning 94–99 face 14–15 face-to-face communication 10–12 factivity 55 fairness 203–206 faking 26 feature based theory 11 fiction 158–161 flouting 78 functional Magnet Resonance Imaging (fMRI) 19–21 gaze 18 generalizing term 18 greeting 24 Gricean Pragmatics 63 guess 80, 82 Guilty Knowledge Task (GKT) 19–20 half-truth 28, 129 honesty 13–15 horizontal convention 160 humbug 174 identity-based deception 10 ill-groundedness 112 illocutionary force indicating device 67 immediacy 18, 96 imperative sentence 53–54 implicature – clausal 128, 130–131 – conventional 35–36, 97, 137–140, 201–202 – conversational 35, 114, 120–123 – existential 121

   253

– generalized conversational (GCI) 35–36, 122, 134–135 – particularized conversational (PCI) 35–36, 122, 134–135, 191–192 – proper 115 – relative 214 – scalar 116, 129–130, 142, 193 – working out schema 116 impliciture 37, 140–147 imprecision 148–152 independence 95 indexicalism 39 indexical pragmatics 63, 84 indicating 187 inference to stereotype 27 insincerity 72–73 insinuation, insinuating 27, 105, 132 insult 109 intention 68 – fictive intention 160 – intention to deceive 103, 104–106, 176 intercultural pragmatics 17, 236 interface approach 47–59 Internet 10–12 interrogative sentence 50–51, 215 intuitive content 37 I-principle 128, 131–132 irony 15–16, 77, 106, 150, 161–167 – echo theory of irony 162–164 – pretence theory of irony 164–167 joke 15, 25, 30, 50, 78 justification 80–81 justification-commitment 80 knowledge lie 108 Knowledge Norm of Assertion (KNA) 225–232 Levinson’s heuristics 127–128 lexical diversity 19 lie 41–47 lie, lying 26, 75, 102–155 – benevolent 152–153 – by untruthful quotation 195 – indirect 26–27, 126 – in order to protect the private sphere 153 – in self-defence 153

254   

   Index

– other-serving 11–12 – self-serving 11–12 – strong lie 103–104, 199 – weak lie 103–104, 199 linguistification 235 loose use 78–79 lügen 41 make-believe 159–160 maxims – maxim of Manner 29, 34, 128, 131, 158, 170, 187, 202, 219 – maxim of Quality 34, 76–79, 108, 119, 128, 150, 162, 166, 170, 209, 223, 232 – maxim of Quantity 24, 28, 34, 61, 128, 170, 187, 202, 232 – maxim of Relevance 34, 123, 202 media richness theory 11 mental load 17 message 200–202 metalinguistic negation 142, 201 metaphor 15, 77, 106 mindfucking 181–182 minimalism 38–39 m-intention 99–100, 184 misleading 31, 105–106, 136–137, 153 misrepresentational intent 175–176 missing constituent 140–141 mistake 16, 126 misunderstanding 136 modal particle 60–62 modulation 38 monotonicity 80 Moore’s paradox 77 M-principle 128, 132–133 mutual knowledge condition 114 narrative 50, 159 negative statement 18 Neogriceans 36 neurology 17–21 nondisplaceability 95–96 non-serious action 187 norm 66, 85, 163 Not Misleading Norm of Assertion (NMNA) 232

obfuscation 29 obsessive-compulsive disorder (OCD) 113 occasionalism 39–40 omission 30–31 optative sentence 53–54 opting out 78 ordinary deception 8 orienting reflex 19–20 original utterance 183–210 overspecifity/narrowness 140–141, 144 overstatement (hyperbole) 106, 168–171, 174 paradox of lies 75 performative verb 67, 93 periphrasis 128, 133 perlocution 74–76 perspective dependence 96 persuasion 76, 148, 179 phoniness 175 plausibility 107–108, 180 plausible answer 18 politeness 14–15 politics 12–14 postsemantic pragmatics 63 pragmatic halo 149, 150–151 pragmatic intrusion 37 pragmatic universal 22 precondition faking 28 preparatory rule 67, 74 presemantic pragmatics 63 presupposition 50–51, 55–59, 61–62, 88–91, 137–140 – hearer-oriented principle 58 – speaker-oriented principle 58 presupposition faking 28 pretending 26, 74, 207 – pretending to act 29 – pretending to joke 29 – pretending to lie 29 Principle of Relevance 36 principle of truthful presupposition 139 private sphere 3–10, 153 promise, promising 72 proper implicature 115 proper name 157 proposition 63, 74, 85, 145 propositional content 66–67

Index   

propositional content rule 67 prosocial lie (white lie) 152–154 prototype semantics 33, 42 psychological partitioning 110 psychology 17–21 public belief 89 Q-principle 128–129, 198 quotation 183–210 – direct quotation 185–195 – emphatic quotation 185 – free (in)direct quotation 185 – indirect quotation 185–195 – mixed quotation 185, 195–199 – pure quotation 185 – scare quotation 185, 199–202 Reasonable to Believe Norm of Assertion (RTBNA) 230 reflexive intention 69 regulative rule 73 relative clause 59–62 – appositive 59 – restrictive 59 relative lie 211–233 relativism 211–214 Relevance theory, Relevance theorists 36, 144–145, 150, 163 reliability 225–232 repeatability 96 representing oneself 100 reservation 82 response length 18 responsibility 70 reticence 27–28 retraction 80, 208 rhetorical assertion 92 rhetorical command 92 rhetorical question 92 Sally-Anne-Task 15–16 sarcasm 77 sarcastic repetition 163 saturation 38 say 185–206 scope ambiguity 142 selective depiction 187

   255

self-deception 8, 109–113 selfless assertion 225–232 self reference 18 semantic interpretation 63 semantic object 217 semantic representation 63 semantics-pragmatics distinction 34–40 semantics-pragmatics interface 38, 63, 235 sentence fragment 184 sentence-meaning 63 sentence mood, sentence modality 49, 86, 217 sentence type 33, 47–59 S-family-test 90, 137–138 simulation 161 – of insincerity 165–166 – of sincerity 165–166 sincerity condition 14, 71–74 sincerity rule 67, 74 slack regulator 149 social distance theory 10–11 sociopath 113 speaker-meaning 63 speech-act assignment 40, 48 speech-act fundamendalists 65–71 speech-act traditionalists 65–71 speech-act type 48 standard expectation 164 statement 65–71, 93, 104 supposition 161 taste predicate 151–152, 211–214 tautology 167–168 taxonomy 8, 30–31 teasing 23 telling distances 149–150 telling the time 149 temporal partitioning 110 Temptation Resistance Paradigm 16 tension 113 testimony 225–232 Theory of Mind 15–16 threat lie 24 total signification of an utterance 120–121 translation 188–189 Transmission of Epistemic Properties (TEP) 226–232

256   

   Index

transparency 203–206 trust 154 truth 83–88 truth-commitment 80 truth-conditional pragmatics 38 truth-conditional semantics 83–88 – basic principle of truth-conditional semantics 84 – most certain principle of truth-conditional semantics 83 truth conditions 83–88 truthfulness 78, 83–88, 139, 148–150 truthiness 224 truth-lies paradox 10 truth value 83–88 unclarifiability 180 underdeterminacy 37, 85, 140–147 underspecificity/weakness 140–141, 143–144

understatement 168–171 universality 22–25 unspecified scope 140–141 utterer-implicature 117 vagueness 148–152 verbatim assumption 185–195 vertical convention 160 verum focus 196, 215–219 violation 78 warranting the truth 104, 106 wave pattern 20–21 what is implicated 1, 34–35, 40, 84 what is said 1, 34–35, 40, 84, 123, 147 wh-question 48 wish 48 wollen 206–209 wording theory 188–190