On the Power of Slurring Words and Derogatory Gestures

Home , Ching chong, Chink, Chinky, Colored, Honky, OK gesture

Charged Expressions: On the Power of Slurring Words and Derogatory Gestures

Ralph DiFranco, Ph.D.

University of Connecticut, 2016

Slurs have the striking power to promulgate prejudice. Standard semantic and pragmatic theories fail to explain how this works. Drawing on embodied cognition research, I show that metaphorical slurs, descriptive slurs, and slurs that imitate their targets are effective means of transmitting prejudice because they are vehicles for prompting hearers to form mental images that depict targets in unflattering ways or to simulate experiential states such as negative feelings for targets. However, slurs are a heterogeneous group, and there may be no one mechanism by which slurs harm their targets. Some perpetrate a visceral kind of harm – they shock and offend hearers – while others sully hearers with objectionable imagery. Thus, a pluralistic account is needed.

Although recent philosophical work on pejoratives has focused exclusively on words, derogation is a broader phenomenon that often constitutively involves various forms of non- verbal communication. This dissertation leads the way into uncharted territory by offering an account of the rhetorical power of iconic derogatory gestures and other non-verbal pejoratives that derogate by virtue of some iconic resemblance to their targets. Like many slurs, iconic derogatory gestures are designed to sully recipients with objectionable imagery. I also address ethical issues concerning the use of pejoratives. For instance, I show that the use of slurs for a powerful majority group by a vulnerable minority may be a morally valuable activity.

Charged Expressions: On the Power of Slurring Words and Derogatory Gestures

Ralph DiFranco

B.A., Cleveland State University, 2008

M.A., Texas Tech University, 2010

A Dissertation

Submitted in Partial Fulfillment of

the Requirements for the Degree of

Doctor of Philosophy

at the

University of Connecticut

2016 iii

APPROVAL PAGE

Doctor of Philosophy Dissertation

Charged Expressions: On the Power of Slurring Words and Derogatory Gestures

Presented by

Ralph Christopher DiFranco, B.A., M.A.

Major Advisor______Mitchell Green

Associate Advisor______Dorit Bar-On

Associate Advisor______William Lycan

University of Connecticut

2016

Acknowledgements

I owe a deep debt of gratitude to Mitchell Green, who has been a generous, nurturing, and patient supervisor for the past five years. I have grown both as a philosopher and a person as a result of working with him. I began work on this dissertation at the University of Virginia, where

I was Mitch’s student before we both moved to UConn. I am indebted to faculty and graduate students at UVA who provided many stimulating discussions regarding my project when it was in its earliest stages of development. Prior to coming to UVA, I worked with Christopher Hom, who deserves special thanks for introducing me to the topic of slurs in a seminar he taught at

Texas Tech University. I have been fortunate to discuss my work in numerous meetings with my committee members, Dorit Bar-On and William Lycan, who have pushed me to think about my project in ways I had not considered previously. Audience members at various colloquia at

UConn over the years have also benefitted me with helpful discussion and challenging questions.

I am especially grateful to Daniel Silvermint, Suzy Killmister, Hallie Liberto, David Ripley, and

Michael Lynch, all of whom graciously read my work and provided insightful feedback.

Content from the first two Chapters of this dissertation appears in an encyclopedia entry,

“Pejorative Language,” which I published in the Internet Encyclopedia of Philosophy. I am grateful to Anthony Neal, my commentator at the American Philosophical Association’s 2015

Central Division meeting, where I presented an early, abbreviated version of Chapter 3, and to

Luvell Anderson, my commentator at the 2016 Pacific Division meeting, where I presented a version of Chapter 6, as well as audience members at both meetings. In 2015, I published a version of Chapter 4 in Thought: A Journal of Philosophy, and Chapters 5 and 6 are currently under review at journals. I have revised all three Chapters in light of extensive referee comments.

Lastly, I am thankful for family and friends who have provided encouragement and support throughout my career as a graduate student.

Table of Contents

Introduction______1

Chapter One: What’s Special about Charged Language?______4

1. A Taxonomy of Expressive Behavior______4

1.1. Pejorative Language______8

1.2. Approbatives______12

1.3. Exclamations and Swears______13

2. Theoretical Benchmarks______14

2.1. Derogatory Power______14

2.1.1. Quantifying the Harmful Effects of Slurs______15

2.1.2. Qualitative Effects of Slurs______16

2.2. Truth-Conditional Content______18

2.3. Expressive Autonomy______19

2.4. Embedded Uses______20

2.5. Appropriation______24

2.6. Heterogeneity______25

3. Conclusion and a Preview of Things to Come______25

Chapter Two: On the Failure of Standard Semantic and Pragmatic Theories______27

1. Introduction______27

2. Expressivism______28

3. A Gestural Theory______33

4. Slurs and Truth-Value Gaps______39

5. Presupposition______41

vii

6. Fregean Minimalism______46

7. Implicature Theories______48

8. Perspectivalism______56

9. Inferentialism______59

10. Stereotype Semantics______61

11. Prohibitionism______67

12. Conclusion______70

Chapter Three: Mental Imagery and the Rhetorical Power of Slurs______72

1. Introduction______72

2. Slurs and Embodied Language Processing______73

3. Other Pejoratives and Approbatives______90

4. The Power of Imagery______93

5. Advantages of a Pluralistic Approach Over Standard Theories______100

6. Appropriation______103

7. Conclusion______104

Chapter Four: Do Racists Speak Truly? on the Truth-Conditional Content of Slurs______106

1. Introduction______106

2. Proponents of NC______107

3. Recent Challenges to NC______108

4. Some New Data______114

5. Conclusion______121

Chapter Five: Derogation Without Words: On the Power of Non-Verbal Pejoratives______123

1. Introduction______123

viii

2. Categories of Non-Verbal Expressive Behavior______124

3. Convention, Natural Design, Idiosyncrasy, and Iconicity______132

4. Imagery and the Rhetorical Power of NVPs______136

4.1.The Role of Imagery in Iconic Representation______137

4.2.Iconicity and the Drift to the Arbitrary______144

5. Imagery and Embodiment in Metaphorical Gestures______147

6. The Inadequacy of Standard Semantic and Pragmatic Theories______149

7. Conclusion______155

Chapter Six: Appropriate Slurs______157

1. Slurs and the Varieties of Social Groups______159

2. Relational Goods and Group Identities______165

2.1. Humorous Derogation______168

2.2. On the Significance of Vulnerability______170

3. Pro Tanto Wrongs ______174

4. Conclusion______179

References______181

Introduction

This dissertation contributes to a growing body of literature on slurring words. In addition to engaging with previous work, it explores uncharted territory by offering a theory of the power of non-verbal pejoratives, and by addressing ethical questions surrounding the use of slurs that have been heretofore unanswered in the philosophical literature.

Before we begin, a disclaimer: while this dissertation mentions many pejorative words, it does not contain any derogatory uses of pejoratives. All occurrences of slurs in this dissertation should be read as metalinguistic. Although we may feel sullied just by reading and interpreting these words, I believe that theorizing about them requires considering examples in which they appear in full, and not in an abbreviated or censored form.

Chapter 1 provides a taxonomy of pejorative language and specifies a set of explanatory desiderata that successful theory of these phenomena should meet. Chapter 2 surveys the philosophical literature on slurs and evaluates previous theories on how well they meet the desiderata specified in Chapter 1.

Chapter 3 motivates an original account of the derogatory power of pejorative words.

Why do metaphorical slurs (e.g., referring to garlic as ‘Italian perfume’), descriptive slurs (e.g.,

‘slanty-eyed’), and imitative slurs (e.g., ‘ching chong’) have the striking power to promulgate bigotry? What accounts for the fact that sometimes merely overhearing utterances of these words makes listeners feel sullied? Drawing on embodied cognition research, I show that slurs like these are effective vehicles for spreading prejudice because they prompt hearers to form

1 unflattering mental images of their targets. Research suggests that, in general, imagery can influence one’s thoughts, emotions, and behavior through a process that Tamar Gendler calls

“imaginative contagion.” Calling an East Asian person ‘slanty-eyed’ invites hearers to form a racist mental image that may lead them to view the target as strange looking (and perhaps even subhuman). However, I do not treat slurs as a uniform, homogenous class. While the embodied approach extends to a large portion of this class, I argue that no one specific mechanism accounts for the power of all slurs.

Chapter 4 challenges the widely accepted view that the truth-conditional content of slurs is identical to that of their neutral counterpart terms. On this view, the literal, truth-conditional content of, e.g., ‘Jew’ and ‘kike’ is identical. I argue that the standard view does not hold for a variety of figurative slurs (‘Jewish American Princess’), descriptive slurs (‘slanty-eyed’,

‘redskin’), as well as some imitative slurs, such as ‘ching chong’, which does not seem to encode any truth-conditional content. These data suggest that slurs do not have uniform truth-conditions.

Non-verbal pejoratives are the focus of Chapter 5. By pulling the skin near one’s eyes outward to give them a “slanted” appearance, one can invite observers to mentally form a racist image that depicts East Asian people as having strange and exotic facial features. Other derogatory impersonations constitutively involve external props. These include “Blackface” minstrel shows (popular in the U.S. from the mid-nineteenth until the early twentieth century) in which actors would darken their skin with makeup and sing songs designed to caricature African

Americans. Chapter 5 accounts for how the iconic features of Blackface performances, including unflattering vocal impersonations of African Americans, contribute to their derogatory power.

Chapter 6 addresses moral and political questions surrounding the use of slurs that have yet to be explored in the philosophical literature. Would it be wrong for an Indigenous group in

2 the U.S. to slur white Americans? Theorists tend to assume that derogation on the basis of characteristics like ethnicity, race, or nationality is, in principle, morally objectionable. Plausibly, slurring is morally wrong to the extent that it promulgates objectionable attitudes or undermines the objective social standing of innocent people. However, the use of a slur needn’t give rise to wrongs like these. I argue that derogation with slurs is a morally valuable activity in three sets of cases. In the first, slurs are vehicles for protesting groups that manifest bad-making features, including white supremacist groups. In the second, slurs are vehicles for constructing a group identity. For instance, an Indigenous group may slur a racial or ethnic majority group to distance themselves from the majority, and thereby reinforce a unique ethnic identity and avoid assimilation. Finally, some slurs are vehicles for a humorous form of derogation that encourages speakers and targets to view themselves in a less serious light and with a certain detachment.

‘Shabbes goy’, e.g., is a humorous slur that Jewish speakers use for gentiles who help them do work when religious restrictions prohibit Jews from working. This slur marks non-Jews as “the other,” yet it does so in a playful way, by gingerly ridiculing them as people who do menial tasks for Jews on the Sabbath. Moreover, the term is often just as much a vehicle for making light of the laws that restrict work as it is for mocking outsiders.

Chapter 1: What’s Special about Charged Language?

1. A Taxonomy of Expressive Behavior

This chapter describes a category of linguistic phenomena that I call ‘charged language’ and specifies a set of explanatory desiderata that a theory of the rhetorical power of these phenomena should meet. I begin with a few general observations about the analysandum. The phrase ‘charged language’ may refer to expressive vehicles, e.g., charged words and phrases.

Uses of charged language are a type of expressive behavior. I shall understand expressive behavior broadly as any behavior designed by nature, convention, or an individual to display or manifest certain internal states, or to produce a certain effect on targeted recipients. I will concentrate on a specific set of expressive vehicles whose conventional role is to motivate attitudes in observes and prompt them to behave in various ways.1 As we will see, however, the distinction of expressive vehicles designed by nature and expressive vehicles designed by convention is not a sharp one.

To get clear on the kinds of expressive acts and vehicles I am interested in, it will be useful to give a general taxonomy of expressive behavior. One plausible necessary condition on a behavior’s being expressive is that it is a signal. Green (2007) distinguishes cues from signals.

He defines a cue as “any feature of an entity that conveys information (including misinformation)” (2007, p. 5). And he defines a signal as “any cue that was designed for its

1 Bar-On (2015) distinguishes acts of expression (or “a-expression”) from vehicles agents use to express psychological states. 4 ability to convey the information that it does” (2007, p. 5).2 For instance, as a result of evolution by natural selection, some tree frogs have brightly colored skin, which signals to predators that the frogs are poisonous (Green, 2007, p. 6). We can distinguish behavior that merely manifests or betrays some information from that which signals information.3 While it is plausible that certain facial expressions can signal an internal state of an organism, a galvanic skin response cannot. Although both might indicate the presence of a certain emotional state, only the smile was designed (by biological evolution) to convey such information. Since the galvanic skin response was not designed to inform observers about the presence of an internal state of the organism, it is merely a cue, and not a signal, and thus not an instance of expressive behavior

(Green, 2007, p. 27). Analogously, the use a charged word, such as a slur, may manifest negative attitudes, though manifesting a negative attitude is not sufficient for signaling its presence. So, if it turns out that slurring words are rhetorically powerful in virtue of what they indicate about a speaker’s negative attitudes, we will need to ask the further question of whether such words are signals of negative attitudes or whether they merely manifest negative attitudes. Additionally, to claim that slurs signal the presence of negative attitudes is not to address the question of whether slurs are conventional signals of the attitudes in question or are signals only by virtue of a speaker intentionally making their attitudes manifest by uttering these words. So, e.g., while the phrase ‘philosopher’ is not a conventional vehicle for signaling negative attitudes toward

2 The use of the term ‘design’ here should not be taken to suggest that a behavior counts as expressive only if it was consciously designed to convey something. It is plausible that some expressive behaviors were designed by natural selection or convention to convey what they do, and both kinds of design can be unconscious and unguided (Green 2007, p. 5). 3 According to Freud, some attitudes are kept out of conscious awareness because they generate mental distress, yet they become manifest in slips of the tongue (Freud, 1949). These “Freudian slips” may be designed by the unconscious to give vent to one’s inappropriate attitudes, and in that sense they may signal their presence. Of course, accepting the possibility of signals designed by the unconscious does not commit us to all of the controversial tenets of Freud’s theory. Kendon (2004, p. 8) denies that unconscious behaviors like these can be expressive, though I will not try to settle this issue here. 5 philosophers, speakers may signal a disparaging attitude toward them by uttering ‘philosopher’ with the appropriate communicative intention. In some cases, derogation with a neutral category label depends on paralinguistic features of an utterance (Jeshion, 2013b). We can imagine a speaker complaining about a disorganized conference by saying, ‘This is what happens when philosophers are in charge’ in a contemptuous tone. The speaker denigrates the philosophers in charge, and seemingly academic philosophers generally, though she does not use a conventional slur. Here a word not conventionally designed to move people emotionally is expressive by virtue of the tone in which it is uttered.

Furthermore, we should take care not to confuse the phenomenon of expressiveness from that of self-expression. I shall adopt the characterization of self-expression developed in Green

(2007):

Where A is an agent and B a cognitive, affective, or experiential state of a sort to which A can have introspective access, A expresses her B if and only if A is in state B, and some action or behavior of A’s both shows and signals her B. (p. 43)

Not all cases of expressive behavior are cases of self-expression: “[o]wing to an accident a person’s face might be disfigured in such a way that it seems to be locked into a sneer” even when the person does not feel contempt (Green, 2007, p. 40). The “sneer” does not show the person’s contempt as long as they lack the relevant attitudes. But because the facial configuration exhibited by the person in question was designed (by natural selection) to convey feelings of contempt, and would commonly be used by individuals to express their contempt, it counts as being expressive of contempt (Green, 2007, p. 40). Similarly, while acting in a play, a non- method actor might utter words designed to convey anger when she is not actually angry. In this case, the actor’s utterance would be expressive of anger, but it would not be a case of self- expression because ex hypothesi it would not show the actor’s anger. Likewise, while slurs and 6 other charged words may be expressive of certain negative (or positive attitudes), uses of such words needn’t be instances of self-expression.

We are now in a position to give a general taxonomy of charged language. It will be useful to divide up the phenomena into two exhaustive categories: the conventional and the non- conventional. I shall use the term ‘conventional’ to refer to vehicles, such as words and phrases, that are widely used by members of a particular group, and whose use is governed by norms.4 In general, conventional behaviors needn’t become conventional by explicit agreement.5

Some have argued that moral and aesthetic terms (such as ‘virtuous’, ‘evil’, ‘beautiful’, and ‘ugly’) make up one category of charged language. These include emotivists, such as Ayer

(1936), who held that moral terms function to persuade others to adopt an attitude of disapproval or approval toward the action in question. Simon Blackburn defends a similar view: 6

When people express themselves in the normative terms of ethics and morals, they are voicing practical attitudes and emotions. They may be doing other things as well: inviting or insisting on others sharing those attitudes or emotions, or prescribing ways to behave, or demanding conformity to ways of behaving. (Blackburn, 2006, p. 148)

Pejoratives, including slurs and other insulting terms, which are words and phrases designed by convention or individuals to denigrate or harm their targets, make up another category of charged language. Parallel to pejoratives are approbatives, i.e., terms of praise. Finally, there are exclamations and swears (e.g., ‘Damn!’). Non-linguistic signs, such as gestures, may also be

“charged” in various ways. One can give a “thumbs up” gesture in order to laud someone.

4 The conventional/non-conventional distinction also cuts across non-verbal expressive vehicles, including gestures. I discuss non-verbal phenomena in more detail in Chapter 5. 5 Some theorists, e.g., Lewis (1969), argue that conventions must arise as solutions to coordination problems. I will not offer a theory of convention here. 6 Ayer held that moral judgments do not express propositions (and so they are neither true nor false). In contrast, Blackburn (2006) allows that moral and aesthetic judgments may be truth apt (see also Gibbard, 1990). 7

Pejorative gesture, such as the “middle finger,” are non-verbal vehicles for derogation. There are also vocal intonations conventionally designed to slur particular groups (e.g., one might speak in a caricatured “African American” accent to ridicule speakers of African American English.)7

In contrast, some non-conventional expressive vehicles are species-general and result from biological evolution, while others were designed by individuals. Green (2007) gives the example of the character Rosamond Vincy from George Elliot’s Middlemarch, who cranes her neck whenever she is feeling obstinate. This gesture was not designed by natural selection to signal one’s feeling of obstinacy, but it could be designed by Rosamond to do so. If observers came to associate Rosamond’s behavior with her emotional life, and she became aware of this, she might begin consciously to contort her neck with the purpose of making manifest her feeling of obstinacy to her audience, and thereby making it perceptible to them (Green, 2007, p. 143). In that case, her behavior could signal her obstinate feelings. As we will see in Chapter 5, grammaticalized expressions such as ‘Yuck!’ and ‘Ouch!’ fall in the middle of a continuum ranging from conventional to natural vehicles (Bar-On, 2015).

1.1. Pejorative Language

Let us now examine each of the various kinds of expressive behavior in more detail.

Following Hom (2010), I will distinguish conventional pejorative words from terms that may carry a negative connotation on many occasions of use, but were not designed by convention to convey anything beyond their neutral, truth-conditional meaning. These include ‘felon’ and

‘diseased’.8 Pejoratives include slurring words, which are designed by convention to derogate people qua members of certain groups, such as a racial or gender group (Jeshion, 2013b, p. 232;

7 Here I concentrate on linguistic phenomena. I give a more thorough taxonomy of non-verbal expressive behaviors in Chapter 5. 8 Notice, however, that over time these terms could become conventionally designed to disparage particular groups or individuals. 8

Hom, 2010, p. 165; Anderson & Lepore, 2013a, p. 25; Bianchi, 2014, p. 12). We can distinguish ostensible targets of a slur, i.e., individuals to whom the slur is applied on a particular occasion of use, from the target group of the slur (Jay, 2009). Often the utterance of a slur need not have any ostensible target (I discuss this distinction in more detail in Chapter 6). Some theorists distinguish slurring words from “particularistic pejoratives,” including ‘asshole’ and ‘jerk’, which are conventionally designed to disparage people on the basis of behavior (Saka, 2007, p.

148; see also Hay, 2013). One alleged difference is that slurs have neutral counterparts, whereas particularistic pejoratives do not (Hay, 2013, p. 456). A neutral counterpart of a slur s is a non- evaluative term that refers to s’s target group, e.g., ‘Italian’ is a neutral counterpart to ‘wop’

(Hornsby, 2001). Compare ‘asshole’, which is conventionally applied to individuals who behave in ways that are (allegedly) inconsiderate, selfish, etc., with the slur ‘wop’, which conventionally denigrates anyone of Italian ancestry, regardless of how they behave. However, Ashwell (2016) argues that having a possible neutral counterpart is not a necessary condition on a word’s being a slur. On her view, potential candidates for a neutral counterpart for the term ‘slut’, such as

‘woman who has more sex than is appropriate’, are not neutral.

One reason for thinking that the distinction of slurs from particularistic pejoratives is not a sharp one is that slurs may target people on the basis of a variety of demographics, some of which combine group-classification and behavior. Slurs based on sexual orientation, or religious or political affiliation challenge the notion that slurs are easily distinguishable from other pejoratives. Religion-based slurs, such as ‘Yid’ and ‘Bible thumper’, are “applied not just on the basis of a kind of abstract group-membership, but for practicing…a certain religion,” and this is not clearly separable from one’s behavior (Jeshion, 2013b, p. 236). Also, slurs that target people on the basis of sexual orientation do not clearly fall on either side of the standard dichotomy. If

Card (1995) is correct, a person’s sexual orientation is neither an abstract, immutable feature that is completely independent of their behavior, nor is it something one freely chooses. Sexual orientation is a complicated feature that may include a self-identification component, a behavioral component, and perhaps other kinds of group-membership that do not constitutively involve any agency whatsoever. Moreover, some pejoratives are designed to be applied to members of a proper subset of a group who exhibit certain character traits or behavior (Nunberg forthcoming). The term ‘dick’, e.g., is applied exclusively to men; however, it is not standardly applied to all men, but only to those who behave in ways deemed overly aggressive or insulting.

As Ashwell (2016) observers, ‘slut’ is a conventional vehicle for disparaging women on the basis of both gender and sexual behavior. Finally, some slurs are applied to proper subsets of a group that allegedly exemplify a set of stereotypical features, including ‘Uncle Tom’, ‘Tiger Mom’, and ‘Jewish American Princess’ (Jeshion, 2013b, p. 235). Thus, it seems that we cannot draw a sharp distinction between slurs and behavior- or character-based pejoratives.

It is possible to create idiosyncratic pejoratives. One could design a rather benign, affectionate nickname like ‘Mr. Stench’ for a friend in order to facilitate the sort of mild, gentle derision that goes on between friends. However, idiosyncratic pejoratives may be highly disparaging. A slur that starts out as idiosyncratic may become a conventional means derogating a group over time (I discuss the conventionalization of slurs in more detail in Chapter 3). Also, there may be borderline cases between the idiosyncratic and the conventionalized (I return to this issue in Chapter 5).

When interpreted as a nouns ‘slur’ and ‘insult’ refer to vehicles for denigrating people on the basis of certain characteristics. But when interpreted as a verb, ‘slur’ and ‘insult’ refer to the action of denigrating people by means of uttering the appropriate word (Anderson & Lepore,

2013b, p. 351). ‘Asshole’, for instance, is an insulting term which was designed by convention to be predicated of targets (such as ‘You are an asshole!’) to insult them (i.e., to offend the target or cause “hurt feelings”).

In the right circumstances just about any behavior can serve to insult(verb) a person, but not all insulting behavior constitutively involves the use of an insult(noun). In other words, one needn’t use a conventional vehicle for offending, denigrating, etc., in order to bring about these effects. One could insult by means of conversational implicature (‘Do you think it would be possible for you to take out the trash?’) or by uttering a non-pejorative expression in a sarcastic tone (‘Nice shirt!’).9 As we have seen, in some circumstances, a neutral term may be used to slur a group, as when someone utters a classifying term, like ‘Chinese’ or ‘gay’, in a contemptuous or dismissive tone (Jeshion, 2013b).

One may feel insulted by the use of a non-pejorative word because of one’s background beliefs. Someone who is skeptical of the value of higher education may take it as an insult if he were called an ‘academic’, and it may be insulting to a working class individual to be called

‘affluent’; a Nietzschean might take it as an insult if she were described as having ‘good

Christian values’; and it might be insulting for a Randian to be called ‘altruistic’. Similarly, although the handshake was not designed by convention to offend, an attempt to shake someone’s hand may be offensive in cultures where greetings do not involve physical contact.

Although, as I have argued, we cannot clearly distinguish slurs from other conventional insults at the level of vehicles, we can draw a fairly clear distinction between slurring and insulting at the level of acts. In general, insulting someone is a matter of causing a particular type of negative psychological state in one’s target, such as making them feel shocked, upset, or

9 Of course, one needn’t use language – or for that matter, do anything at all – to insult. In some cases, omissions, such as failing to hold a door open for someone or not making eye contact during a conversation, can be insulting. 11 offended. Agents who insult may or may not be culpable for doing so (I will not give an account of what it is to culpably insult someone here). A woman who wears clothing designed for men, e.g., may insult people who think that women should satisfy traditional gender norms, yet the woman in question need not be culpable for having done so. Whereas being (merely) insulted is simply a matter of being shocked, upset or offended (or something of the sort), one needn’t be in any psychological state like these in order to be slurred (Hom, 2012, p. 397). Consider, e.g., the use of the slur ‘nigger’ during a Ku Klux Klan rally. Even if no one at the rally is insulted or offended by the use of the term (because all of the participants are racist), African Americans are nevertheless slurred by the use of this term. In other words, there are contexts in which use of

‘nigger’ is derogatory independently of whether anyone involved is offended by (or knows about) its use.

1.2. Approbatives

Parallel to pejorative terms are words and phrases designed by convention or individuals to praise or evince approval. Following Hom (2008), I shall call such terms approbatives.10

Among the positively “charged” words designed by convention are terms for physical appearance (‘knockout’, ‘stunner’), status or achievement (‘superstar’), and so on. Additionally, speakers can introduce idiosyncratic approbatives, i.e., words or phrases designed by individuals to praise. Some approbatives are applied on the basis of gender: whereas ‘goddess’ is applied exclusively to women, ‘stud’ is applied exclusively to men. Others are gender-neutral.

1.3. Exclamations and Swears

Some words are conventionally designed simply to express a heightened emotional state, which may be positive or negative (Potts, 2007). These include exclamations (‘Oops!’) and

10 Bach (2014) refers to these terms as ‘laudatives’. 12 swears (‘Fuck!’, ‘Shit!’, ‘Damn!’).11 Swears are unique among other expressive words and phrases in that they exhibit a great deal of versatility in how they can be positioned syntactically.

As Hom (2012, p. 384) observes, some swears can be used to modify non-expressive content:

(1) Fuck! The managing partner fired John.12

And, as Hom (2012) notes, swears can appear in the adjective or adverb position:

(2) The fucking managing partner fired John!

(3) Mary and John are a fucking couple!

(4) The managing partner fucking fired John!

Swears can also be used to modify pejorative terms to increase their expressive capacity, and in some cases, their level of offensiveness or derogatoriness. An utterance of (5) would typically convey something harsher than an utterance of the same sentence without the swear word:

(5) You fucking idiot!

Notice that swears are often vehicles for expressing positive emotions, as in (6):

(6) The Huskies fucking won!

Additionally, some swears can be used as pejoratives, as in (7) and (8):

(7) Fuck off!

(8) Eat shit!

11 Bach (2014) calls these “expressive intensifiers.” 12 Hom (2010, p. 165) provides additional examples of pejorative exclamations. 13

In Chapter 3, I offer an account of the rhetorical power of slurs, approbatives, and insults like (7) and (8), though I will bracket uses of swears and exclamations, such as those in (1)-(6), and moral terms.

2. Theoretical Benchmarks

One aim of this dissertation is to give a theory about how charged words such as slurs can promulgate derogatory attitudes and sully hearers (or readers). A successful theory of how pejorative words accomplish these aims should meet several explanatory desiderata.13 I will focus on six central features in need of explanation: derogatory power, truth-conditional content, expressive autonomy, embedded uses, appropriation, and heterogeneity. I shall explicate each feature in turn. This will set the stage for Chapter 2, in which I survey the logical space of theories of expressive behavior and evaluate each on how well it meets these benchmarks.

2.1. Derogatory Power

Pejoratives are vehicles for promoting disparaging ways of thinking about their targets.

Calling someone ‘loser’, e.g., is a way of pressuring listeners to disapprove of them, and treat them as undesirable, damaged, inferior, etc. Additionally, (non-sarcastically) calling someone

‘stunner’ is a way of pressuring listeners to regard them as exceptionally attractive. So, one feature of charged words that requires an explanation is their power to influence speakers and their listeners to adopt certain negative or positive attitudes toward, and to motivate them to behave in corresponding ways toward, their targets. I will examine both quantitative and qualitative measures of slurs’ derogatory power.

13 I offer a set of explanatory desiderata for a theory of non-verbal pejoratives in Chapter 5. As we will see, the list of desiderata I give here and the one I give in Chapter 5 overlap somewhat, though they are not identical. 14

2.1.1. Quantifying the Harmful Effects of Slurs

A substantial body of literature in the social sciences is devoted to studying the effects of slurs on listeners. Greenberg & Pyszczynski (1985) report that participants who viewed a staged debate between a Black and a White debater, and then overheard someone posing as a participant comment, ‘There’s no way that nigger won the debate’ evaluated the Black debater as less skilled than the White debater. Evaluations in conditions where participants overheard the slur were harsher than in conditions in which the fake participant did not use a slur and instead said,

‘There’s no way that pro debater won the debate’, referring to the Black debater. Goodman et al.

(2008) found that overhearing a slur for homosexuals applied to a gay male leader of a work group led to a “perceived diminution of his leadership abilities” among group members (p. 549).

Finally, Gadon & Johnson (2009) examined the effects of overhearing ‘shrink’ applied to a psychologist. Participants who watch a video of a therapy session, and then overheard the psychologist in the video referred to as a ‘shrink’ evaluated him as having less expertise and as being less considerate, and were less willing to seek help from the psychologist observed, than in cases where a professional label (‘psychologist’) or no label was used.14

Empirical data suggest that automatic reactions to slurs vary depending on whether the hearer self-identifies as part of the target group. Carnaghi & Maass (2007) found that priming heterosexuals with anti-gay slurs (such as ‘fairy’ and ‘fag’) automatically elicited negatively valenced associations with homosexuality (in comparison with neutral labels ‘gay’ and

‘homosexual’, which did not). They also report that anti-gay slurs did not negatively affect participants who self-identified as LGBT, whose reactions to slurs did not differ from their

14 Gadon & Johnson reported no significant differences between conditions where participants heard the professional label and conditions where no label was used (in the latter cases, the psychologist was referred to by the name ‘Mr. Smith’). Interestingly, participants in this study rated the commenter who used the term ‘shrink’ as being more judgmental, tactless and cold than in conditions where a professional label or no label was used. 15 reactions to neutral labels. In a study by Fasoli et al. (2015), participants who self-identified as heterosexual were exposed to the Italian anti-gay slur, ‘frocio’, and were then asked how much money they would allocate to various social welfare programs. Fasoli and colleagues report that exposure to the slur led subjects to allocate less money to a program they took to be relevant to homosexuals (HIV prevention), than to a program they took to be relevant to heterosexuals

(sterility prevention).15

2.1.2 Qualitative Effects of Slurs

In some cases, merely overhearing a speaker utter a pejorative is enough to make non- prejudiced listeners feel tarnished.16 Slurs in particular tend to get under our skin in this way.

Camp (2013) calls this feature ‘complicity’ (however, as I argue in Chapter 2, the notion of complicity does not capture slurs’ sullying effects). The qualitative dimension of derogation may be divided into two categories. Slurs characteristically provoke an immediate, visceral response in both targets and bystanders. We (i.e., non-prejudiced competent listeners) may feel shocked or offended as a result overhearing a prohibited term for a group (Anderson & Lepore, 2013a).

Additionally, slurs tend to have a more long-lasting sullying effect, which may be divided into two forms. For targets, simply overhearing the use of a slur may (but need not) be demoralizing.

During a 2013 Australian Football League match, player Adam Goodes, who is of indigenous

Australian ancestry, reported feeling gutted after being called an ‘ape’ by a fan (Windley, 2013).

15 For more quantitative data on the effects of anti-gay slurs, see Carnaghi & Maass, (2007) and Goodman et al. (2008). 16 This is not to suggest that all uses of slurs are likely to sully hearers. Most theorists accept that slurs are inoffensive and non-derogatory when quoted (though Anderson and Lepore are dissenters to this view). And appropriated uses of epithets by members of the target group may be inoffensive, though the (in)offensiveness of such uses are often contested (Brontsema, 2004). I will have more to say about appropriation in the current Chapter as well as in Chapter 3. Additionally, Hom (2008) argues that there are non-derogatory, non-appropriated (NDNA) uses of epithets. Hom holds that, e.g., a speaker who asserts ‘Institutions that treat Chinese as chinks are racist’ literally says something both non-derogatory and true (p. 425). I address Hom’s purported NDNA uses in Chapter 2. 16

Secondly, non-targets may be sullied when overhearing slurs because they are led to entertain certain unpleasant thoughts or images that they would rather not entertain.17 Sullying is not just a characteristic effect of overhearing others use slurs; one may sully oneself with a slur. Even uttering a slur to oneself internally, i.e., within the privacy of one’s own mind, can lead to unpleasant thoughts or images. Of course, listeners might feel sullied without actually being sullied, and vice-versa. In other words, overhearing a slur could make us feel tarnished, yet not motivate us to think and behave as racists do. Conversely, we should acknowledge the possibility that slurs might influence us without giving us an introspectively accessible experience.

A number of theorists observe that slurs vary in their derogatory power – some tend to be more insulting and derogatory than others (Saka, 2007). Seemingly, the use of ‘limey’ is much less derogatory toward the English than the use of ‘wetback’ is toward Hispanic people.

Likewise, co-referential slurs can vary in derogatory power (Anderson & Lepore, 2013b). The slurs ‘guinea’ and ‘wop’ both target Italians, yet many regard the former as much more offensive and derogatory than the latter. While some theorists reject derogatory variation (see, e.g.,

Jeshion, 2013b), empirical data support claims of derogatory variation among different pejoratives. In a study by Kirkland et al. (1987), subjects read a transcript of a jury trial in which a Black attorney defended a White client. In one condition, participants overheard a speaker posing as a participant refer to the lawyer using the slur ‘nigger’. In these cases, participants evaluated the lawyer negatively (even when they were outwardly disturbed by the use of the slur). However, overhearing the lawyer referred to in an ethnically neutral pejorative way (with

‘shyster’) did not lead to negative evaluations. Additionally, Carnaghi et al. (2011) found that exposure to Italian anti-gay slurs led subjects who identified as heterosexual to distance

17 Camp (2013, p. 343) points to a similar feature, “cognitive complicity,” which she claims occurs when overhearing a slur automatically activates a derogatory “perspective” in listeners. I discuss Camp’s view in more detail in Chapters 2 and 3. 17 themselves from homosexuals and emphasize a masculine identity on an assessment of their gender identity, whereas slurs for regional identity (namely, ‘terrone’, which targets residents of southern Italy) did not have these effects.

2.2. Truth-Conditional Content

Gibbard (2003) suggests that the notion of a thick ethical concept, due to Williams

(1985), can shed light on the meaning of slurs. In comparison with thin ethical concepts (such as right and wrong, just and unjust), thick ethical concepts contain both evaluative and descriptive content. Paradigm examples include cruel, cowardly, and unchaste. For Williams, these terms not only play a role in prescribing and motivating action; they also purport to describe how things are. To say that a person is cruel, e.g., is to say that they bring about suffering, and they are morally wrong for doing so. According to Gibbard,

[r]acial epithets may sometimes work this way: where the local population stems from different far parts of the world, classification by ancestry can be factual and descriptive, but, alas, the terms people use for this are often denigrating. Nonracists can recognize things people say as truths objectionably couched. (2003, p. 300)

As we will see, while Gibbard’s claim that slurring statements express truths that are

“objectionably couched” is controversial,18 it does seem that many slurs classify their targets. A speaker who calls an Italian person ‘spic’ has not merely said something offensive and derogatory – he has also made a factual error: he has classified his target incorrectly. Similarly, the insult ‘moron’ appears to both ascribe a low level of intelligence to its targets and evaluate them negatively for it. Additionally, as the following example illustrates, some swear words contain descriptive content:

18 I raise objections to Gibbard’s view, which is a variant of a theory I call ‘neutral counterpart theory’, in Chapter 4. 18

(9) A: Tom fucked Jerry for the first time last week.

B: No, they fucked for the second last week; the first time was two months ago.

Also consider the following example (Hom, 2010, p. 170):

(10) Random fucking is risky behavior.

There appears to be genuine disagreement between A and B in (9), and someone who asserts

(10) has surely made a claim capable of being true or false.

One question that a theory of slurs ought to address is whether the derogation accomplished by slurs crucially relies on slurs encoding truth-conditional content. In Chapter 4, I point to words that I call “imitative slurs” (such as ‘ching chong’), which crucially involve an iconic element, to motivate the view that slurs needn’t encode truth-conditional content in order to denigrate their targets. Additionally, some slurs appear to contribute truth-conditional content that goes beyond classifying someone’s group membership (e.g., ‘slanty-eyed’). A successful theory of pejoratives needs to explain (or explain away) their apparent descriptive, truth- conditional features.

2.3. Expressive Autonomy

An expressive behavior is autonomous to the extent that what it conveys and the characteristic effects it has are independent of the background beliefs of the speakers who perform the behavior. Slurs appear to exhibit derogatory autonomy, i.e., the derogatory capacity of slurs is independent of the derogatory attitudes (or lack thereof) of the speakers who use them

(Hom, 2008, p. 426). As Anderson and Lepore (2013a, p. 33) observe, a racist who asserts

‘Chinks are smarter than the rest of us’, “intending to compliment or to announce affection” has

19 still used the slur in a patently offensive manner. Likewise, a competent speaker who knows full well that ‘kike’ is a term of abuse for Jews could not stipulate a non-derogatory meaning by uttering, “What’s wrong with saying that kikes are smart? By ‘kike’, I just mean Jews, and Jews are smart, aren’t they?” (Saka, 2007, p. 148).

2.4. Embedded Uses

Potts (2007) observes that most pejoratives appear to exhibit nondisplaceability: except in certain cases where a speaker is directly quoting someone else, uses of pejoratives seem to convey negative content when uttered. In other words, pejoratives seem to carry expressive content that tends to “scope out” of embedding. Consider (11)-(14). Suppose that Sue uttered

(11) and (12) is a report on her utterance; also suppose that Eric uttered (13) and another speaker has attempted to report on his utterance with (14):19

(11) That asshole Steve is on time today.

(12) Sue said that that asshole Steve is on time today.

(13) A bitch ran for President of the United States in 2016.

(14) Eric said that a bitch ran for President of the United States in 2016.

As long as the occurrences of ‘asshole’ and ‘bitch’ are not read as implicitly metalinguistic (with a change in intonation or an accompanying gesture indicating that the speaker wishes to distance herself from any negative feelings toward Steve), listeners will interpret the speaker of (12) as making a disparaging remark about Steve, even if the speaker is merely attempting to report on

Sue’s utterance. Likewise, as Anderson and Lepore (2013a, p. 29) note, it would be difficult to use (14) to give a neutral (non-sexist) report on Eric’s offensive claim (see also Saka, 2007, p.

19 Potts (2007) and Anderson and Lepore (2013a) offer similar examples. 20

122). Unless a metalinguistic reading is available for the occurrence of ‘bitch’, anyone who utters (14) in an attempt to report on Eric’s utterance of (13) risks making an offensive claim about women.

Potts holds that one way in which pejoratives are nondisplaceable is that they always tell us about the current utterance situation (2007, p. 169). Consider (15):

(15) That bastard Kresge was late for work yesterday (#But he’s no bastard today, because he was on time)

Despite the fact that ‘bastard’ is within the scope of a tense operator in (15), it would be implausible to read the speaker as claiming that she disliked Kresge only in the past, as the defective parenthetical (indicated by the hash sign) illustrates (Potts, 2007, p. 171).

However, not all pejoratives behave the same way when embedded. Consider (16)-(19):

(16) If Steve doesn’t finish his report by the end of the week, he’s fucked (but I suspect he’ll finish on time.)

(17) Suppose our new employee, Steve, is a bastard (On the other hand, maybe he’ll be nice)

(18) Steve is not a bastard (I think he’s a good guy).

(19) Steve used to be a real motherfucker in law school (but I like him much better now).

A speaker who utters (16)-(19) needn’t be taken as having made a disparaging claim about Steve.

This is because the occurrences of ‘fucked’, ‘bastard’, and ‘motherfucker’ in (16)-(19) appear to be “narrow-scoping” (Hom, 2012, p. 387). Thus, at least some embedded uses of pejoratives seem not to commit the speaker to an offensive claim (compare the non-defective parentheticals in these cases with the defective one in (15)).

The most offensive pejoratives, however, appear to always scope out. The slurs in (20) and (21) are just as offensive and derogatory when uttered as part of a supposition or embedded in a conditional sentence as when they are used in predicative assertions:20

(20) If the guys standing at the end of my driveway are spics, I’ll tell them to leave (#Fortunately, there is no such thing as a spic, since no one is inferior for being Hispanic)

(21) Suppose the next job applicant is a nigger. (#Of course that won’t happen, since no one is inferior for being Black.)

Some slurs appear to take wide scope relative to all truth-conditional operators, including negation. Consider the following explicit attempt to reject a racist claim:

(22) It is not the case that Ben is a kike; he’s not Jewish!

(22) fails to undermine the derogatoriness of the slur ‘kike’. Seemingly, the trouble is that it only disavows a derogatory way of thinking and talking about Ben, and so it cannot be used to reject a racist attitude toward Jews in general (Camp, 2013).

A successful theory of pejoratives must explain the behavior of embedded pejorative words and phrases, and more specifically, must account for the fact that some pejoratives behave differently within the scope of truth-conditional and intensional operators. A successful theory must also resolve the apparent tension between the putative descriptive features of slurs, and their behavior under embedding.

The nondisplaceability of pejoratives is one feature that sets them apart from moral predicates (Hay, 2013). In contrast with slurs, moral and aesthetic terms (‘good’, ‘bad’,

‘beautiful’, ‘ugly’) may be uttered without making a moral or aesthetic judgment. Compare (20)

20 McCready (2010) gives similar data in demonstrating the wide-scoping behavior of slurs. 22 and (21), which would say something racist in nearly all contexts of use, with (24)-(28), which would not express a vegetarian sentiment in most contexts of use:

(24) If eating meat is immoral, it should be a crime. (But it’s not immoral.)

(25) Suppose that eating meat is immoral. (Of course, I don’t really think it is.)

(26) George thinks that eating meat is immoral. (But I disagree.)

(27) Is eating meat immoral?

(28) It’s not the case that eating meat is immoral.

Notice that nondisplaceability is a feature exhibited only by certain words embedded within the scope of a truth-conditional or intensional operator. Thus, it would be a category mistake to claim that a non-linguistic expressive behavior is nondisplaceable. But there is a parallel feature exhibited by non-linguistic expressives. It is difficult to perform the middle finger gesture without ‘flipping off’ one’s audience. In most cases, we would be hard-pressed to find a performance of the middle finger gesture that does not incur an offense, even in cases when one is merely attempting to demonstrate the gesture. Similarly, it would be very difficult to wear a Klan uniform without incurring a great deal of offense.21 In comparison, if one wanted to present a highly insulting term like ‘asshole’ to an audience, and at the same time shield oneself from the potentially negative consequences of using it, one could utter the term in a metalinguistic context (e.g., by quoting it); but there is no analogue for non-linguistic pejorative displays.

21 Things get more complicated, however, when slurring occurs within the context of a theatrical performance. It might be argued that when a non-racist actor plays a racist character who utters a slur, the actor is isolated from derogation. On the other hand, there might be cases in which we’d want to say that the actor uttering the slur did something racist and derogatory, and this might depend on features of the play (or film or television program) in which she is acting. Alternatively, we may be inclined to attribute a racist action to the director or author. Similar considerations apply to non-linguistic slurring behavior. 23

2.5. Appropriation

Some kinds of expressive behavior are systematically used to accomplish aims other than those for which they were designed. I shall use the term ‘appropriation’ broadly to refer to the various systematic ways in which agents repurpose expressive behaviors.

For certain slurs, the target group has taken over the term in order to transform its meaning to lessen or eliminate its derogatory force. This is one variety of appropriation known as linguistic reclamation (Brontsema, 2004). The term ‘queer’ is a paradigm case. Although ‘queer’ was once used to derogate those who engaged in sexual behavior deemed abnormal, it now has virtually no derogatory force due to efforts by LGBT speakers to appropriate the term. As a result, non-prejudiced speakers can now use ‘queer’ in a variety of contexts (e.g., in ‘queer studies program’ and ‘queer theory’) without thereby derogating LGBT people. In contrast, the slur ‘nigger’ (often marked by an alternative spelling ‘nigga’) has been appropriated more narrowly to just the target group, and is often used as a means of expressing camaraderie between group members (Saka, 2007, p. 145). Except for certain rare exceptions, only African

Americans can use the term to refer to each other without derogating one another. Appropriated uses of ‘nigger’ are common in comedic performances and satire. The use of ‘nigger’ in a comedy bit designed to mock and criticize racism needn’t commit the speaker to hateful attitudes toward African Americans (Richard, 2008, p. 12). Additionally, non-linguistic slurring behavior can be appropriated by target groups. An oppressed group could appropriate displays or gestures used by their oppressors in order to change or eliminate their derogatory power.

Finally, we can appropriate swears and gestures to express only mild disapprobation in talking to friends. In such cases, the speaker’s conversational aim is more teasing and jocular

24 than insulting. A successful theory of these phenomena will need to account for their various appropriated uses.

2.6. Heterogeneity

Theorists should take into account the fact that different pejoratives fall into different grammatical and lexical categories. Pejoratives may function grammatically as noun phrases

(‘dago’), adjectives (‘nappy-headed’), or verbs (‘to Jew’, ‘to gyp’); some are figurative (‘Jewish

American Princess’), while others (such as ‘kike’) are not (Jeshion, 2013b, pp. 310-11). Further, as we have seen, different pejoratives seem to disparage in different ways. Some slurs appear to rely on stereotypes (Saka, 2007; Hom, 2008; Jeshion, 2013b), yet others may simply be abbreviations of a neutral label for the target group (‘Abo’ for Aboriginals, ‘Jap’ for Japanese, etc.). Some describe perceptible features of their targets (‘redskin’, ‘slanty-eyed’), while others imitate their targets (‘ching chong’). Do all slurs function in precisely the same way, despite their apparent differences? Is a unified theory of slurs tenable? A successful theory should explain (or explain away) slurs apparent heterogeneity.

3. Conclusion and a Preview of Things to Come

To summarize, a successful theory of charged language should answer the following questions:

(i) What accounts for the derogatory power of pejorative words, including their power to sully us and their more immediate, visceral, less cognitively demanding effects?

(ii) Do all slurs encode truth-conditional content, and if so, is this content identical to the content that their corresponding neutral classifying terms encode?

(iii) Expressive autonomy: why are uses of slurs are derogatory even when speakers do not intend for them to be?

(iv) Why does embedding fail to isolate a speaker from offensiveness?

(v) How does the appropriation of a slur undermine its offensiveness?

(vi) Should we be optimistic about a unified theory of pejoratives or is this class of linguistic phenomena so diverse that a unified theory is untenable?

This dissertation focuses on a limited set of pejoratives and approbatives, and brackets other charged words, including exclamations, and moral and aesthetic terms. Drawing on embodied cognition research, I show in Chapter 3 that many slurs promulgate derogatory attitudes by prompting hearers to form mental images and simulate experiential states.

Descriptive slurs such as ‘slanty-eyed’ prompt hearers to mentally form a caricatured image of the target. Empirical data indicate that images can have real psychological and behavioral effects outside of the context of an episode of imagining. This is due to “imaginative contagion,” whereby mental imagery (when not quarantined) bleeds into one’s attitudes (Gendler, 2006).

However, we needn’t hold that all slurs harm in the same way. As I argue in Chapter 2, standard approaches fail by treating slurs as a unified class. Some slurs are more effective vehicles for the visceral sort of derogation described above. These slurs tend to shock hearers rather than sully them, and this does not require depicting a target in an unflattering way. Slurs that were once imagistic can lose their imagery over time in a process analogous to what Tomasello (2008) calls the “drift to the arbitrary,” which is the transformation of iconic gestures into arbitrary gestures.

In many cases, speakers and hearers are unaware of the original basis for finding a formerly imagistic slur objectionable, yet a taboo on the term remains. Thus, slurs that have drifted to the arbitrary may still be conventional vehicles for shocking and offending hearers, though they no longer sully hearers with vivid imagery.

Chapter 2: On the Failure of Standard Semantic and Pragmatic Theories

1. Introduction

This chapter assesses various accounts of charged language, focusing in particular on theories of pejorative and approbative words. Traditional theories divide into two general categories: semantic accounts, which take the distinctive features of charged words and phrases to be explained at the level of what they semantically encode, and pragmatic theories, which instead look to what a word or phrase is used to imply, pragmatically presuppose, etc.22 I also consider a deflationary account, due to Anderson and Lepore (2013a), according to which slurs do not express any specifically derogatory content and are merely “prohibited words,” as well as a perspectival account, due to Camp (2013), which does not clearly fit on either side of semantic/pragmatic distinction. I evaluate each theory on how well it meets the benchmarks articulated in the previous chapter. Historically, there has been a presumption in favor pragmatic accounts of charged language, and so I will begin by evaluating the various pragmatic approaches. I argue that traditional approaches fail to meet the explanatory desiderata specified in Chapter 1, and while the prohibitionist account and perspectival views provide valuable insights, they too are inadequate. One crucial explanatory desideratum that the current menu of alternatives fails to meet is an account of the rhetorical power of slurs and approbatives.

22 This is not to assume that Grice’s distinction of what an utterance literally means from what it implies is entirely unproblematic, or to make any other controversial assumption about the relationship between semantics and pragmatics. I discuss worries about the Gricean picture in Section 7 below. For now, all I am stipulating is that there is some plausible notion of literal meaning on which an utterance’s literal meaning is distinct from what a speaker means by it. I do so only in order to find a starting point for exploring the logical space of various theories of slurs. 27

2. Expressivism

One popular theory of slurs is a descendent of the metaethical view known as expressivism. According to the version of expressivism developed by Ayer (1936), moral and aesthetic statements do not express propositions capable of being true or false, and merely serve to express and endorse the speaker’s own moral sentiments. For Ayer, an assertion of ‘Stealing is wrong’, e.g., merely evinces the speaker’s disapproval of stealing. On a naïve expressivist theory of pejoratives, derogatory statements containing pejoratives do not express propositions capable of being true or false – they merely express a non-cognitive attitude (such as disapproval) directed at the target group. A theory of this sort is well suited to explain the behavior of slurs under embedding. However, it will have difficulty accounting for their descriptive features. As noted previously, a speaker who calls an Italian person ‘spic’ has (seemingly) made a classificatory error. If slurs lack descriptive content, and merely serve to express non-cognitive attitudes, then it is unclear how they could classify their targets.

Saka (2007) offers a hybrid expressivist theory of slurs, according to which slurs contain both expressive and descriptive content (see also Kaplan, 2004). Saka denies that there is a single belief or proposition expressed by slurring statements such as ‘Nietzsche was a kraut’. Rather, such statements express an attitude complex, which includes (i) the pure belief that Nietzsche was German, and (ii) a negative cognitive-affective state (such as hostility) toward Germans

(Saka, 2007, p. 143). Saka’s hybrid theory accounts for the descriptive, truth-conditional features of pejoratives. More recently, Jeshion (2013b) has defended an expressivist theory that builds upon the work of Saka and Kaplan. She offers a three-component expressivist semantics for slurs. First, slurs have the same truth-conditions as their neutral counterparts. Second, slurs have an “expressive” component: the use of a slur expresses contempt for members of a target group

G “on account of their being in G or on account of their possessing a G-defining property g”

(2013b, p. 316-7). Jeshion also claims that in using slurs, speakers represent their targets as worthy of contempt, but do not say or assert that their targets are contemptible (2013b, p. 318).

On her view, contempt is both an “affectively- and normatively-guided moral attitude,” which is

“held toward those one regards as inferior as persons” (2013b, p. 318, emphasis in original).

Finally, slurs have an “identifying component”: in using a slur, a speaker “takes a property that he believes someone to possess and semantically encodes that it is the, or a, defining feature of the target’s identity” (Jeshion, 2013b, p. 318).

Notice that merely expressing contempt for S is insufficient for representing S as worthy of contempt. Suppose that Smith is given a promotion over Jones. Jones may express genuine contempt for Smith by uttering, ‘That damn Smith got the promotion I was up for!’ or by uttering, ‘Smith got the promotion’ in a contemptuous intonation. Some theorists hold that contempt is a “globalist” emotion in that it takes the whole individual as its target (Bell, 2013).

However, Ben-Ze’ev (2001) holds that contempt can be localized to specific features of an individual, and so one can feel contempt for another on the basis of their looks, e.g., yet this does not entail feeling contempt for the whole individual. Jeshion seems to assume globalism (see her

2013b, p. 242). I will not try to settle the question of whether contempt is a globalist emotion or potentially compartmentalized; regardless, it is not clear why harboring contempt for a target requires thinking of them as worthy of that attitude. Jones could consistently follow up her utterance about Smith by sincerely claiming that her contempt for Smith is unfounded, that she thinks Smith deserved the promotion, and that she does not regard the opinions of others who do not feel contempt for Smith as defective. We have no reason to think that contempt cannot be a- rational in this way, such that it diverges from our considered judgments about a person. Thus,

29 much more needs to be said to show that contempt is itself a normatively-guided attitude.

Moreover, even when one attributes group membership to someone in a contemptuous way, one needn’t be taken as having represented the target as worthy of contempt because of their group membership. In yelling “Hey, red!” while scowling at someone, one needn’t represent the target as contemptible because of his red hair (Hom, 2008, p. 418). Of course, we can represent people as worthy of contempt by ascribing properties to them that make them contemptible. One could represent Italians as contemptible because for their group membership by saying something like, ‘Italians are characteristically violent, despicable people’.23 However, since Jeshion denies that speakers who use slurs say or assert that their targets are contemptible because of their group membership, it is difficult to see how slurs represent targets negatively on her approach.

Another problem is that the “identifying component” in Jeshion’s account seems unmotivated and opens her account up to a number of counterexamples. As noted previously, slurs can target people on the basis of a variety of different features, including immigrant status, nationality, or occupation. It is doubtful that any of these are “defining features” of a person’s identity. We can imagine a bigot who uses ‘kraut’ and holds that being German is a defining feature of a German person’s identity, yet speakers who use ‘kraut’ to denigrate Germans are not thereby committed to a view like this. Similarly, there is no reason to think that speakers who use

‘yuppie’ commit themselves to the view that a person’s occupation and socio-economic status are defining features of their identity. Further, as Jeshion herself acknowledges, ‘faggot’ targets people on the basis of a feature that is not obviously essential, namely sexual orientation (see also Ashwell, 2016). To assume that slurs for sexual orientation groups target people on the basis

23 Notice that imagery may also do explanatory work here; the right kind of image could represent someone as worthy of contempt (or pity, disgust, etc.). I am indebted to Mitchell Green for helpful discussion on this point. 30 of a feature that is essential to their identity is to assume a conception of sexual orientation that is, at best, highly controversial. Regardless, it is not clear that bigots who use ‘faggot’ are committed to holding that sexual orientation is an essential or defining feature of one’s identity.

Standard objections to metaethical expressivism raise challenges to the various expressivist accounts of pejoratives considered so far. According to one such objection, the so- called Frege-Geach problem, one can utter a sentence containing a moral predicate (such as

‘good’, ‘evil’, ‘right’, or ‘wrong’) as the antecedent or consequent of a conditional sentence without making a moral judgment. Expressivists about moral terms are unable to account for the sameness of content in both asserted and unasserted contexts, so the objection goes. For example, as Geach (1965) observed, the following is a valid argument:

(1) If tormenting the cat is wrong, then getting your brother to do it is wrong.

(2) Tormenting the cat is wrong.

(3) Therefore, getting your brother to do it is wrong.

If, as the metaethical expressivist claims, ‘wrong’ merely expresses a speaker’s disapproval, then it is a mystery how the term ‘wrong’ could carry the same content in (2) and when embedded in the antecedent of the conditional sentence in (1), given that (2) expresses a moral judgment while

(1) does not. Hom (2010, p. 171) argues persuasively that expressivist theories of pejoratives face a similar challenge. While some pejoratives scope out of truth-conditional operators, some appear to functionally embed. Consider the following argument:

(4) If George fucked up his presentation, he’ll be fired.

(5) George fucked up his presentation.

(6) Therefore, he’ll be fired.

In order for this argument to be valid, the pejorative phrase ‘fucked up’ must have the same semantic content in (4) and (5), despite the fact that (4) does not express a negative attitude about

George, while (5) does. It is difficult to see how the naïve expressivist theory could account for this. Although hybrid theories have the potential to explain the preservation of content between

(4) and (5), they will have difficulty accounting for the fact that no negative attitude about

George is expressed by (4).

Another standard objection to expressivist accounts of slurs is that the non-cognitive attitudes they posit are too underspecified to account for derogatory variation, e.g., ‘kraut’ is less derogatory than ‘nigger’ (Hom, 2008). Do all pejoratives express something like contempt or do the negative attitudes differ for each term? Saka holds that apparent derogatory variation among slurs can be explained away by pointing to the historical circumstances that led to their introduction and sustain their derogatory power (Saka, 2007, p. 148). Jeshion (2013b) takes a similar approach in attempting to explain away derogatory variation:

An utterance of ‘Nigger’ will typically cause considerably worse psychological and social damage than an utterance of ‘honkey’ in part because the former occurs against the background of current widespread racism, history of slavery, and historical civil rights struggles for African-Americans, and nothing comparable for Caucasians. (p. 322)

It is not clear, on Jeshion’s view, how much of a slur’s offensiveness is supposed to be explained by the history of the term’s use and other background conditions. The appeal to historical context here is illicit if the derogatoriness of slurs is to be explained by an attitude complex expressed by speakers who use the term. For once external institutions are appealed to in order to explain the derogatory features of slurs, it would appear that the posited attitude complex no longer does any explanatory work, and so the expressivist rejoinder to those who posit variation seems self- undermining. Moreover, appealing to historical conditions does not explain away apparent

32 variation of co-referential slurs.24 For instance, the slurs in each of the follow pairs –

‘darkie’/‘nigger’ and ‘guinea’/‘wop’ – each seem to vary in their offensiveness.25 Further, slurs seem to vary not only in their degree of offensiveness and derogatory power, but also in the kinds of derogation they are meant to accomplish. It is doubtful that jocular slurs and relatively mild slurs such as ‘yuppie’ are expressive of any degree of contempt (Camp, 2013, pp. 332-3; I return to this issue in Chapters 3 and 6).

Finally, expressivists need to do more to explain how the expression of negative attitudes relates to the rhetorical power of slurs. In particular, they need to specify a notion of expression by means of slurs that makes it clear how the expression of hostility (or contempt, etc.) toward a target could motivate listeners to feel similarly. As Camp (2013) observes, if expressivism is correct, “the hearer should be able to dismiss the speaker’s feelings as just her problem” (p. 333, emphasis in original).

3. A Gestural Theory

Hornsby (2001) offers a gestural theory of slurs:

It is as if someone who used, say, the word ‘nigger’ had made a particular gesture while uttering the word’s neutral counterpart. An aspect of the word’s meaning is to be thought of as if it were communicated by means of this (posited) gesture. The gesture is made, ineludibly, in the course of speaking, and is thus to be explicated…in illocutionary terms. (p. 140)

According to Hornsby, the gestural content of a slur cannot be captured in terms of a proposition or thought it conveys. Rather, “the commitments incurred by someone who makes the gesture are commitments to targeted emotional attitudes” (p. 140). Because gestures are inherently physical events with perceptible properties, their expressive power cannot be fully explained with words.

24 Anderson & Lepore (2013b) raise a similar objection to the semantic theory developed by Hom (2008). 25 I am indebted to David Ripley for a helpful discussion on this point. 33

When one sneers or draws an index finger horizontally across one’s throat, e.g., one expresses a psychological state that cannot be fully captured linguistically. Pejorative gestures disparage and put down their targets, but attempts to gloss what these gestures communicate in terms of propositional speaker meaning (e.g., ‘I dislike you’ or something of the sort) do not capture the way they do so.26 Slurs appear to exhibit a similar sort of “ineffability”: as Potts (2007) observes, competent speakers are unable to articulate the derogatory contents of slurs using descriptive terms (p. 166). Attempts to paraphrase the slur ‘nigger’ as ‘inferior for being African American’ or ‘African American and despicable because of it’, e.g., seem to fail to capture the raw, negative attitudes expressed by the use of the term.

Hornsby’s gestural theory accounts for the embedded behavior of slurs (see Chapter 1,

Section 2.4.), and could plausibly extend to cover approbative terms. Hornsby might argue that to call someone a ‘gentleman’, e.g., is to call them a male while simultaneously expressing admiration for them via a positively charged gesture, such as tipping one’s hat. Also, a gestural theory could potentially account for variation in derogatory capacity of different slurs. So, e.g., the slur ‘limey’ might convey something like a “thumbs down” while uttering ‘English’, whereas

‘chink’ might convey something more derogatory, like a “middle finger” accompanied by an utterance of ‘Chinese’.

While Hornsby’s suggestion that derogatory gestures are more fundamental vehicles for expression than derogatory words seems plausible, there are two major lacunae in her gestural theory: she is not clear on what the posited gestures are supposed to be, nor does she say anything specific about how a slur and its corresponding gesture are related. Thus, evaluating her theory requires filling in some details. To begin, it will be helpful to distinguish gestures which are products of convention (e.g., the “middle finger”, “thumbs down”, and “throat slash”) from

26 However, gestures can become conventionalized over time. 34 those which are products of natural selection (these include configurations of an organism’s face and features of an organism’s voice designed by natural selection to convey basic emotions).27

This distinction allows for two versions of Hornsby’s proposal. On the first, slurs convey pejorative gestures designed by convention; on the second, slurs convey pejorative gestures designed by natural selection. Let us consider each version in turn.

Suppose Hornsby’s view is that calling someone ‘nigger’ is like calling them ‘Black’ while performing a conventional gesture, say, the “throat slash”. Hom (2008) worries that this proposal fails to explain the derogatory nature of the slur. To slur someone is, inter alia, to denigrate them by means of communicating something negative about their group. Notice that to call a person ‘Black’ while directing a “throat slash” at them is not necessarily to derogate them on the basis of their race, since the racial classification might be completely incidental to the speaker’s expression of a negative attitude via the gesture.28 Hom suggests that any accompanying conventional gesture Hornsby might posit for a given slur would face a similar problem, so the worry goes.

However, pace Hom, gestures may be designed by convention to derogate particular individuals on the basis of group membership. Perhaps the “Nazi salute” is a conventional means of denigrating Jews and other groups historically targeted by the Nazis (even if it was not originally designed for that purpose). In that case, Hornsby could argue that ‘kike’, e.g., has as its gestural content something like the content expressed by a Nazi salute. This suggestion would yield a promising theory if for each slur there were a corresponding gesture conventionally

27 As I explain in Chapter 5, these gestures fall on a continuum that runs from the fully conventionalized to the natural. 28 One might think that to perform a gesture like the “throat slash” while simultaneously uttering a neutral classifying term for a certain group G is to conversationally imply that members of G are despicable or inferior because of being members of G. Unfortunately, this strategy is not available to Hornsby, since she takes the content of slurs to be a non-propositional attitude. Potts (2007) offers a non-propositional conventional implicature theory, which I discuss below. 35 designed to derogate the target group that determines the gesture’s pejorative content. Yet this is doubtful. Moreover, Hornsby does not offer any account of the derogatory capacity of conventional gestures. If her aim is to explain the derogatory power of slurs by positing conventional gestures, she still needs to provide an answer to the question, how do linguistic and social conventions imbue gestures with the power to derogate?29

An alternative approach for Hornsby is to posit gestures designed by natural selection to express a negative emotion. Humans and other primates evolved the capacity to convey basic emotions by means of certain facial expressions and vocal patterns (including anger, fear, surprise, disgust, sadness, and contempt). If Hornsby is positing gestures designed by nature to convey basic emotions, she would avoid the burden of having to explain how gestures designed by convention get their capacity to express negative attitudes, since we already have a plausible evolutionary story about how basic emotions can be communicated by means of facial configurations and vocal intonations. Consider the sneer, which is a natural signal (in the sense discussed in Chapter 1) of anger. On the present version of the gestural theory, to call someone

‘kike’, e.g., is to ascribe the property of being Jewish to them using a negative tone of voice or while sneering (or producing some other natural threat signal).

As we have seen, however, it is not clear that the use of a slur is derogatory in virtue of expressing a negative emotion. Further, as Camp (2013) observes, pejoratives need not express any of the plausible candidates among the basic emotions that might serve as their emotional content. ‘Yuppie’ seems more jocular than contemptuous. Also consider ‘champagne socialist’, which is a vehicle for accusing wealthy people who purport to be socialists of failing to live up to their supposed convictions, yet it does not appear to do so by means of expressing an emotion.

Another problematic case is ‘Shabbes goy’ (sometimes written as ‘Shabbos goy’ or ‘Shabbat

29 See Chapter 5 for my own view on the power of derogatory gestures. 36 goy’), which targets gentiles who help Jewish people do work on the Sabbath when orthodox law prohibit Jews form working.30 It is not clear what the corresponding gesture or facial expression that provides the affective content of the slur is in this case. Among facial expressions, the closest analogue may be something like a wry smile, yet even this expression does not seem to capture the sort of fine-grained, mild derogation for which this slur is a vehicle.

We have seen that both approaches to filling in the details about Hornsby’s posited gestures (i.e., understanding them as conventional or natural) yield unsatisfactory results. Even if a satisfactory way of spelling out what the posited gestures are is forthcoming, there remains another lacuna in Hornsby’s gestural theory: she still owes an account of how slurs and their corresponding gestures are related. She might say that the derogatory use of a slur elides a pejorative gesture. Before developing this approach, it will be helpful to consider examples of the linguistic phenomenon known as ellipsis. Familiar examples of ellipsis include sluicing in

(7), verb-phrase ellipsis in (8), and noun-phrase ellipsis in (9):

(7) Alexis went out, but I don’t know where.

(8) Alexis can drive and Shelia can, too.

(9) Alexis won five medals, and Mary won six.

The complete thoughts expressed in (7)-(9) could be glossed as Alexis went out, but I don’t know where she went in (7), Alexis can drive and Shelia can drive, too in (8), and Alexis won five medals, and Mary won six medals in (9). There are also fragmented answers to questions, as in

(10):

(10) Q: Who can dunk a basketball? A: Ted.

30 I discuss this example in more detail in Chapter 6. 37

In these cases, ellipsis works anaphorically – the elided material in each of the above cases is parasitic on antecedent material.31

Perhaps Hornsby holds that derogatory uses of slurs involve an analogous phenomenon.

The idea is this: the use of a slur elides a gesture directed at the slur’s target. In uttering the slur, the speaker actually performs a pejorative gesture, which lacks perceptible properties. It is difficult to discern what relationship a slur would have to its corresponding gesture on this approach, however. When ellipsis occurs with language, the elided material is either anaphoric, or receives partial phonological expression by the speaker. The notion of anaphora is inapplicable to Hornsby’s posited gestures, and it is not the case that the speaker who uses a slur produces a “partial gesture.” In general, it is difficult to make sense of the notion of a gesture

(i.e., an expressive limb movement or bodily orientation) that lacks perceptible properties.

Another way of filling in the details is to say that the derogatory use of a slur implies, hints at, suggests (or commits the speaker in some other way to) the performance of a pejorative gesture.

But this approach is also problematic; gestures are constituted by bodily movements, and the notion of implying a bodily movement seems to involve a category mistake. Of course, we can abbreviate a gesture in such a way that recipients can infer or imagine what the additional movement would be, but that would require that the gesture receive at least partial expression in the performer’s bodily movements, and speakers often do not even perform partial or abbreviated gestures when using slurs (at least, they needn’t).

Let us take stock. I have examined two lacunae in Hornsby’s gestural theory: she owes a more precise account of (i) the posited gestures themselves – what they look like and how they succeed in denigrating their targets – and (ii) how slurs and their corresponding gestures are

31 Here I have merely produced a few uncontroversial examples of ellipsis – this is by no means a full taxonomy. For a more thorough lay of the land, see Johnson (2008). 38 related. I then explored several ways of filling in the details. Hornsby’s posited gestures could be due to conventional design or design by natural selection, and there are two ways to characterize the relationship between slurs and their corresponding gestures: the use of a slur either elides a gesture or implies a gesture. Each way of filling in the details either led to an implausible result or replaced old mysteries with new ones.

4. Slurs and Truth-Value Gaps

Seemingly, some derogatory utterances cannot be rejected by means of negation

(Richard, 2008). Suppose George utters (11):

(11) Jill shares an office with a chink.

According to Richard, even if we suppose that Jill’s officemate is not Chinese, it is not clear that

George’s utterance of (11) would be aptly characterized false. Jill couldn’t call George out on his racism by responding with

(12) That’s false, since my officemate is French and not Chinese.

For, according to Richard, this would amount to construing George’s error as a mere mistake in identifying a person’s race, and leave his racist claim unchallenged. Further, (12) seems to amount to an implicit endorsement of the racist attitudes expressed by (11). Similarly, George couldn’t take back his racist statement by adding

(13) My mistake. What I said was false, since your officemate isn’t Chinese.

Richard (2008) argues that sentences containing derogatory uses of slurs are not truth-apt.

He denies that slurring speech is false on the grounds that applying the term ‘false’ to an utterance like George’s would be “to say that he made a particular kind of error…which can be

39 corrected merely by…judicious use of negation” (p. 25). But examples like (11)-(13) suggest that it cannot. Richard also denies that derogatory statements containing slurs can be true. He acknowledges that predicating a slur of someone entails classifying them as a member of a particular group, but he denies that correct classification suffices for truth. For instance, Richard holds that anti-Semites classify a person as Jewish by calling them ‘kike’, but when a speaker slurs a Jewish person with ‘kike’ they have not simply classified them as Jewish nor have they merely expressed an affective state (like hatred or contempt); rather, they have misrepresented the target as being despicable for being Jewish. According to Richard, we cannot endorse the racial classification as true without also endorsing the representation as accurate. For “[w]hatever truth belongs to a classification is truth it inherited from the thought expressed in making it,” and the thought expressed by the anti-Semite through his slurring performance is the mistaken belief that Jews are despicable for being Jewish (Richard, 2008, p. 24).32

Although Richard’s view could potentially make sense of the behavior of slurs under embedding, he does not offer a positive theory of how slurs represent their targets. He could hold that slurs represent targets negatively by expressing certain propositions about them. Or maybe the relevant sort of representation is imagistic. Perhaps he thinks that hearing a slur puts an unflattering image of the target group in the minds of listeners.33 Richard offers no help here.

Instead, he is interested only in establishing that there are numerous statements – including derogatory utterances containing slurs – that have a determinate content, yet are not truth-apt

(others include applications of vague predicates to borderline cases and statements that give rise

32 Thus, Richard rejects a conventional implicature account of slurs, according to which an assertion of ‘George is a limey’ is literally true (so long as George is English), but conventionally implies the false and derogatory proposition that George is inferior for being English. As far as I can tell, Richard doesn’t offer an argument against the conventional implicature theory of slurs. I give Gricean theories their due in Section 7 below. 33 I defend a view like this in Chapter 3. 40 to liar paradoxes). The upshot is that there are certain aspects to meaning – such as the representational features of a sentence – that cannot be explicated in terms of truth conditions.

Also, nothing in Richard’s view helps us understand how slurs have the power to motivate listeners to think derogatory thoughts about their targets or sully listeners.

Finally, there are reasons to be doubtful of Richard’s claim that slurs always misrepresent their targets.34 While this claim seems plausible in the case of racial slurs, it is not obviously true of all slurring words. Consider ‘fascist’, which is a slur conventionally applied to officials and institutions in an authoritarian political system. On Richard’s view, to call Mussolini and Hitler fascists is to represent them as contemptible for their political affiliation. Presumably this would not be to misrepresent them. Richard might agree, and respond that the concept of truth is not what we should use when we’re evaluating someone’s slurring performance as accurate or inaccurate. In that case, Richard still owes a positive account of how such words can be used to accurately represent their targets. Absent these details, it is difficult to see the motivation for

Richard’s view.

5. Presupposition

A third approach to charged language makes use of the notion of presupposition. In general, to presuppose a proposition P is to take P for granted. There are a variety of ways to do this. For instance, when I desire to eat the last slice of pie, I presuppose that it hasn’t already been eaten; and when I sit in a chair, I presuppose that it can support my weight. My focus will be on presupposition in communication: to presuppose a proposition P in this sense is to take P for granted in a way that contrasts with asserting it (Soames, 1989, p. 553). Soames (1989) characterizes the difference between asserting and presupposing as follows: “If assertive

34 Camp (2013) also challenges Richard on this point. 41 utterances of a sentence S are used to assert A and presuppose P, then assertive utterances of more complicated sentences containing S often presuppose P without carrying any commitment to A” (p. 554). For example, in assertively uttering (14), one presupposes that there is a mess and asserts that George is responsible for it:

(14) George is responsible for the mess.

An assertive utterance of the conditional sentence in (15), which embeds (14), carries the same presupposition as (9), but does not commit the speaker to its assertoric content:

(15) If George is responsible for the mess, he should clean it up.35

Theorists standardly distinguish semantic presupposition from pragmatic presupposition.36

According to some theorists, a speaker who utters a sentence s semantically presupposes a proposition P relative to a context c iff P must be true in order for s to express a proposition in c

(Soames, 1989, p. 562). According to Strawson (1950, p. 333), an utterance of (16) generates a semantic presupposition:

(16) This is a fine red one.

This statement presupposes that the demonstrative pronoun, ‘this’, succeeds in picking out some contextually salient object. If it did not, (16) would fail to express a truth-evaluable proposition.

In contrast with semantic presupposition, pragmatic presupposition is defined in terms of the attitudes of speakers and audience members (Stalnaker, 1999, p. 48). On this approach, each

35 Complex constructions do not always inherit presuppositions of sentences they embed. For instance, an utterance of “If the typewriter was broken, then it was Sam who broke it,” does not presuppose that the typewriter was broken, even though its consequent does (Soames, 1989, p. 554). I discuss cases like this in greater detail below. 36 Here one could make finer-grained distinctions concerning different types of communicative presupposition. For a more detailed taxonomy, see Soames (1989). 42 conversation is governed by a conversational record, which includes the common ground, i.e., the background assumptions mutually accepted by participants for the purposes of the conversation. The pragmatic presuppositions of an utterance are the “requirements that [it] place[s] on sets of common background assumptions built up among conversational participants”

(Soames, 1989, p. 556).

The set of propositions that speakers stipulate is subject to change over the course of a conversation. Lewis (1979) gives a general rule of accommodation for presupposition: “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 in to existence at t” (p. 340). Lewis’s idea is that information can be added to (or removed from) the conversational record when necessary in order to forestall presupposition failure and make what is said conversationally acceptable.

One advantage of the pragmatic approach to presupposition over the semantic approach is that the former allows for the requirements that an utterance places on the conversational record to come from a variety of sources, including Gricean mechanisms (such as conventional implicature) and non-conventional pragmatic facts (Soames, 1989, p. 568). Further, according to

Stalnaker (1999), since the pragmatic approach explains facts about what is taken for granted in a conversation by appeal to general principles about rational communication, there is no need to suppose that the requirements that an utterance places on the conversational record are written into its semantics:

The propositions that P and that Q may be related to each other, and to common beliefs and intentions, in such a way that it is hard to think of a reason that anyone would raise the question whether P, or care about its answer, unless he already believed that Q. More generally, it might be that one can make sense of a conversation as a sequence of rational actions only on the assumption that the speaker and his audience share certain presuppositions. If this kind of explanation 43

can be given for the fact that a certain statement tends to require a certain presupposition, then there will be no need to complicate the semantics or the lexicon. (Stalnaker 1999, p. 55)

If statements containing ‘honky’ semantically presuppose that Whites are inferior because of being White (or something of the sort), then such statements would lack truth-values, since presumably, Whites are not inferior because of being White. In that case, speakers could not assert, e.g., ‘Ralph is a honky’. Yet this semantic presupposition analysis of slurs sheds little light on why uses of slurs are offensive. Thus, most contemporary theorists who defend presupposition accounts explain the power of slurs using theoretical resources provided by theories of pragmatic presupposition. According to Schlenker (2007), charged words and phrases

“carry a presupposition…which is indexical (it is evaluated with respect to a context), attitudinal

(it predicates something of the mental state of the agent of that context), and sometimes shiftable

(the context of evaluation need not be the context of the actual utterance)” (p. 237, emphasis in the original). On this view, to use a slur in a derogatory context, e.g., is to attempt to incorporate into the conversational record a certain proposition about members of a particular race, gender, nationality, etc. Consider (17):

(17) Was there a honky on the subway today?

If none of the participants in the conversation dissent, a proposition (or set of such propositions) such as that the speaker and her audience believe that Caucasians are despicable for being

Caucasian, that the speaker and her audience are prepared to treat them such) is incorporated into the conversational record. This approach could be extended to cover approbatives: to call someone a ‘knockout’, e.g., would be to submit the proposition that the speaker and audience believe that the targeted individual is handsome or beautiful (or something of the sort) for

44 inclusion into common ground.

There are several problems with the presupposition approach. First, as Potts (2007), Hom

(2010), and Anderson and Lepore (2013a) observe, presuppositions can be cancelled when embedded in a larger construction, but the offensiveness of slurs cannot. Compare (18) with (19):

(18) Frank believes that John stopped smoking, but John has never smoked.

(19) #Eric said that a nigger is in the white house, but Blacks are not inferior for being Black.37

Ordinarily, an assertion of “John stopped smoking” presupposes that John previously smoked.

When embedded in an indirect report, however, the presupposition can be cancelled, as (18) illustrates. In contrast, (19) appears to convey something derogatory about African Americans, which cannot be cancelled by the right conjunct. If the presupposition account were correct, we would expect (14) to be felicitous, non-derogatory, and inoffensive, yet it is not.

See also the following examples from Potts (2007, p. 170). (20) appears to express something negative toward Kresge despite the embedding:

(20) Sue believes that that bastard Kresge should be fired (#but he’s a good guy).

A similar phenomenon occurs with embedded approbatives:

(21) George believes that that stud Ralph is coming to the party (#He’s not all that).

Also, as Richard (2008) has observed, derogation with slurs needn’t be a rational, cooperative effort between speakers, yet presupposition is:

37 In this case, the occurrence of the slur ‘nigger’ should not be read as implicitly metalinguistic (with a change in intonation or some other gesture indicating that the speaker wishes to distance herself from derogatory attitudes expressed by the use of the term). Likewise with (20) and (21). 45

[a] pretty good rule of thumb is that someone who is using these words is insulting and being hostile to their targets. But there is a rather large gap between doing that and putting something on the conversational record. If I yell ‘Smuck!’ at someone who cuts me off…[a]m I entitled to assume, if you don’t say ‘He’s not a smuck’, that you assume that the person in question is a smuck, or are hostile towards him? Surely not (pp. 21-2).

Since the conversational record needn’t track what participants actually think, allowing a presupposition onto the conversational record does not require actually believing it. Thus, it is difficult to see how pragmatic presupposition can explain slurs’ power to inculcate prejudice.

Finally, in some cases, derogation by means of slurs does not require any common ground whatsoever. For example, a racist might use a slur in foro interno, i.e., in the privacy of her own mind, in the course of what we may call a “derogatory soliloquy” (cf. Williamson,

2009). Imagine someone saying to himself, internally, ‘There are too many wops around here’.

In this case, the speaker is denigrating Italians without attempting to add anything to a conversational record.

6. Fregean Minimalism38

According to Gottlob Frege, there are two aspects to the meaning of a term: its sense (or mode of presentation) and its reference (or what it denotes). Additionally, Frege posited a realm of communication separate from sense and reference. For Frege, a word’s färbung (often translated as ‘coloring’ or ‘shading’) is constituted by the negative or positive psychological states associated with it that play no role in determining the truth-value of utterances that include it. The terms ‘dog’ and ‘cur’, e.g., share the same sense and reference, but the latter tends to carry a negative coloring – something like disgust or contempt for the targeted canine (Frege,

38 I am borrowing this name for a Fregean theory of charged language from Hom (2008, p. 420). 46

1966). Similarly, Frege would argue that (22) and (23) share the same sense and reference, but differ in their coloring:

(22) Mary is English.

(23) Mary is a Limey.

For Frege, both (22) and (23) are true just in case Mary is English. However, for most speakers,

‘English’ is neutral in coloring, while ‘Limey’ is associated with negative feelings for English people.

Although the Fregean approach to pejoratives could account for their behavior when embedded, most contemporary theorists reject it (see, e.g., Hom, 2008; Anderson & Lepore,

2013a). Frege held that “coloring and shading are not objective, and must be evoked by each hearer or reader” (1966, p. 61). Thus, a term’s coloring is not conventional in any sense.

Dummett (1981) diagnoses the problem with positing an essentially subjective realm of meaning:

“[m]eaning, under any theory whatsoever, cannot be in principle subjective, because meaning is a matter of what is conveyed by language” (p. 85, emphasis in original). Given the nature of coloring, Fregeans are committed to holding that the derogatory content of slurs (whatever this turns out to be) is located at the level of subjective psychological states of speakers and listeners.

As a result, Fregeans cannot account for expressive autonomy (see Chapter 1, Section 2.3). For instance, proponents of this view will have trouble explaining why ‘nigger’ can be just as derogatory in the mouth of a racist as it is when uttered by a non-racist (Hom, 2008, p. 421).

In reply, Fregeans might offer a dispositional theory of coloring. Consider an analogy with a dispositional theory of color, according to which a thing is yellow, e.g., if it disposes normal agents in appropriate conditions to have a qualitative experience of yellow. Similarly,

Fregeans might hold that a slur S has a negative coloring to the extent that uttering or hearing S

47 disposes speakers and listeners to have derogatory attitudes toward the target. This approach could generalize to other pejorative terms. Take Frege’s example of ‘cur’. On the revised version of the theory, ‘cur’ has a negative coloring to the extent that competent listeners who hear the term predicated of a dog are disposed to think of the targeted canine as flea-ridden, mangy, and dangerous. Such an account might be promising, but much more would need to be said about how hearing the word disposes listeners to instantiate derogatory attitudes. As it stands, the

Fregean view does little to explain how pejoratives can be so rhetorically powerful. But that does not entail that the theory is hopeless. In some ways, the account I develop in Chapter 3 is a development of the reconstructed Fregean view considered here. Thus, rather than elaborating on what I take to be the right approach in the present chapter, I will defer discussion of it until the next chapter.

7. Implicature Theories

Although Frege’s original theory of coloring does not have many adherents today, he does get credit for observing that speakers can use words like ‘but’, ‘still’ and ‘although’ to mean more than they say. Frege observed that “[a] speaker uses ‘but’ when he wants to hint that what follows is different from what might at first be supposed” (1997, p. 63). Some commentators, including Neale (1999) have argued that Grice’s theory of implicature owes a great debt to

Frege’s theory of coloring.

Grice (1989) also observes that sometimes what a speaker means is not exhausted by what she literally says. Grice posited two kinds of implicature: conversational and conventional.

When a speaker communicates something by means of conversational implicature, she violates

(or makes as if to violate) a conversational norm, such as quantity: provide as much information as is required given the aim of the conversation. The hearer, working on the assumption that the

48 speaker is participating in the conversation cooperatively, then attempts to derive the implicatum

(i.e., what the speaker meant, but did not literally say) based on the words used by the speaker and what conversational norm she has (apparently) violated. In Grice’s classic example, a letter of recommendation for a student, X, reads, “X’s command of English is excellent, and his attendance at tutorials has been regular” (Grice, 1989, p. 33). The most reasonable explanation for the author’s apparent violation of quantity is that she thinks X is a rather bad student, but is reluctant to explicitly say so, since doing so would entail saying something impolite or violating some other norm.39

Sometimes the conventional meaning of a term determines what is implied by a use of it, in addition to determining what is said by it. If a sentence s conventionally implies that Q, then it is possible to find another sentence s*, which is truth-conditionally equivalent to s, yet does not imply that Q (Grice, 1989, p. 58). Consider the sentences ‘Mara is rich and kind’ and ‘Mara is rich, but kind’. For Grice, these two sentences have the same literal truth-conditions (they are true just in case Mara is both rich and kind), but only the latter implies that there is a contrast between being rich and kind (in virtue of the conventional meaning of ‘but’). Conventional implicata are not worked out by the hearer in the way the conversational implicata are.

Several philosophers have defended a Gricean account of charged language. However, few (if any) claim that the derogatory content of slurs is always expressed by means of conversational implicature. First of all, if there is any derogatory content expressed by uses of slurs, it does not need to be worked out by the listener in the way that a conversational implicature is (Hom, 2008, p. 423). Second, conversational implicata are supposed to be

39 This is an example of particularized conversational implicature, which relies more heavily on features of a particular context than generalized implicatures. The latter denotes cases in which “the use of a certain form of words in an utterance would normally (in the absence of special circumstances) carry such-and-such an implicature” (Grice, 1989, p. 37). 49 cancellable, but the derogatory content of a slur is apparently not (Hom, 2008, p. 434).

According to Grice, for any putative conversational implicature P, it will always be possible to explicitly cancel P by adding something like ‘but not P’ or ‘I do not mean to imply that P’. That derogation with slurs is not explicitly cancellable is suggested by the following defective attempt at cancellation: ‘That house is full of kikes, but I don’t mean to disparage Jewish people’.40

Stenner (1981), Whiting (2007, 2013), Williamson (2009), and Lycan (2015) have argued that slurs convey derogatory content by means of conventional implicature. I will take

Williamson’s view (hereafter, the ‘CI account’) as representative. On the CI account, slurs and their neutral counterparts have the same literal meaning, but slurs conventionally imply a proposition that their neutral counterparts do not. Unlike conversational implicata, conventional implicata are not explicitly cancellable. In asserting ‘Franz was a Boche’, the speaker literally says that Franz was German, but conventionally implies that Franz was cruel and despicable (or something of the sort) because he was German (Williamson, 2009).

Hom (2008) argues that “non-derogatory, non-appropriated” (NDNA) uses of slurs pose a challenge to the CI account. As noted in the previous chapter, appropriation occurs when speakers take over a slur for their own group and transform its conventional meaning in order to eliminate its capacity to derogate. According to Hom, some uses of slurs are not appropriated uses (i.e., they’ve retained their conventional meaning), yet they are also not derogatory. These include uses “in pedagogical contexts about racism” (2008, p. 429). Consider, e.g., (24)-(26):

(24) Institutions that treat Chinese people as chinks are morally depraved.

(25) There are no chinks; racists are wrong.

40 Camp (2013, p. 340) and Jeshion (2013a, p. 317) also provide data designed to show that derogation with slurs is non-cancellable. 50

(26) Chinks are (supposedly) despicable because of their race, but Chinese people aren’t.

NDNA uses are allegedly problematic, since the CI account entails that “[d]erogation ought to occur in every context of use for epithets without any means of cancellation” (Hom, 2008, pp.

438-9).

Griceans might reply that Hom’s purported NDNA uses can be read as implicitly metalinguistic, e.g., a Gricean might gloss (24)-(26) as (27)-(29):41

(27) Institutions that treat Chinese in the way that racists treat the people they call ‘chinks’ are morally depraved.

(28) One ought not call people ‘chinks’ (racists are wrong in calling people ‘chinks’)

(29) People called ‘chinks’ are (supposedly) despicable because of their race, but Chinese people are not.

The idea is that in (24)-(26) the epithet is only mentioned as opposed to being used. If no derogatory content is conveyed by these utterances, it is because disguised quotation marks surround the epithet in each sentence, and quotation can isolate a speaker from an epithet’s derogatory implications.

While Griceans have a satisfactory reply to Hom’s objection from NDNA uses, there are other worries about the CI account of slurs. First, it is controversial whether there are such things as conventional implicata. Bach (1999), e.g., argues that putative cases of conventional implicature are actually part of what is said by an utterance. Bach devised the indirect quotation

(IQ) test for conventionally implicated content. Suppose that speaker A has uttered (30), and speaker B has reported on A’s utterance with (31):

41 Hornsby (2001, p. 129) anticipates a strategy like this one. See also Whiting (2013, pp. 370-372). 51

(30) She is wise, but short.

(31) A said that she is wise and short.

According to Bach, since B has left out important information in her indirect report, namely information about the purported contrast between being wise and short, that information must have been part of what was said, as opposed to what was implied, by A’s utterance. Hom (2008) uses Bach’s IQ test to undermine the CI account of slurs. Suppose A uttered (32) and B reported on A’s utterance with (33):

(32) Greg is a spic.

(33) A said that Greg is Hispanic.

According to Hom, since B has misreported A, the derogatory content of the slur must be part of what is said, and so the CI account fails (however, see Lycan, 2015, for a reply).42

A more serious objection is that even if CI theorists are correct in holding that an utterance of ‘Italians are wops’ carries a negative implicature about Italians, more would need to be said in order to explain how implying something negative about Italians could motivate anti-

Italian bigotry. This is not to suggest that propositional attitudes cannot motivate. If the Humean theory of motivation is correct, beliefs coupled with relevant desires (or other conative states) are sufficient for motivation. So, the belief that Italians are inferior plus a desire to treat people the way they ought to be treated could yield discrimination toward Italians. But it is not clear how listeners are supposed to get the relevant beliefs about Italians by means of a conventional implicature. Consider a paradigm case of conventional implicature: a speaker who asserts ‘P but

42 Although Hom makes use of the IQ test in order to motivate his own semantic theory of slurs, it is not clear that the test shows that the derogatory content of slurs is part of their semantic content, since “what is said” by an utterance could refer to pragmatically enriched content, e.g., what Bach (1994) calls “impliciture”. I discuss Hom’s use of Bach’s test in more detail in Chapter 4. 52

Q’ commits herself to a contrast between P and Q by virtue of the conventional meaning of ‘but’.

However, there is no reason to think that bystanders would be inclined to hold that there is some kind of tension between P and Q. Even if terms like ‘but’ are capable of encouraging a kind of implicit acceptance of a tension, this would be very different from the sullying effects of slurs.

For one, we can reject a conventional implicature with a straightforward denial, as when one responds to (30) with ‘there’s no contrast between being wise and short!’ We could attempt the same sort of denial in response to slurs (e.g., responding to a derogatory use of ‘slanty-eyed’ by pointing out that the speaker has implied something racist about the target’s eyes that we should reject), yet such a denial would not address the imagistic effects that a slur such as this is likely to have (I will have more to say about this in Chapter 3).43

This worry generalizes to other developments of the CI account. Lycan (2015), for instance, develops a version of the CI account using his theory of lexical presumption (see

Lycan, 1984). He characterizes lexical presumption as follows: “If a sentence S1 lexically presumes a sentence S2 and S2 is false, then S1 will be heard as deviant in a particular way, viz., as mislexicalized” (Lycan, 2015, p. 5). On this view, the use of a slur lexically presumes a

“negative normative belief, typically a derogatory belief or an attitude of contempt” (Lycan,

2015, p. 6). Given Lycan’s characterization of lexical presumption, the use of a slur would be heard as inappropriate or poorly chosen (over the slur’s neutral counterpart) if the racist beliefs their use presumes are false (which they presumably would be for most slurs). It is not clear, however, what this has to do with slurs’ sullying effects, which often have an imagistic or experiential basis. We have no reason to think that the mere presumption of a belief the hearer takes to be false is likely to lead them to entertain derogatory thoughts and make them feel as

43 Notice that a similar worry applies to presupposition accounts. Explicitly refusing to allow a racist remark into common ground does not address the slur’s imagistic, experiential component. 53 though they have been somehow infected by the use of the term.

Additionally, CI theorists are committed to a claim that I call the ‘neutral counterpart’ thesis (NC), namely that the truth-conditional content of a slur is identical to that of its neutral counterpart. As we will see (in Chapter 4), it is implausible that NC extends to all slurs.

Specifically, ‘slanty-eyed’, ‘Jewish American Princess’, and ‘redskin’ are problematic for the CI account.

Potts (2007) offers a non-propositional version of the CI account. Potts understands pejorative content in terms of expressive indices, which model a speaker’s negative (or positive) attitudes in a conversational context. He offers the following schema for an expressive index:

where a and b are individuals, and I is an interval that represents a’s positive or negative feelings for b in the conversational context. The more narrow the interval, the more intense the feeling.

For instance, if I = [-1, 1], then a is essentially indifferent toward b. If I = [0.8, 1], then a has a highly positive attitude toward b. If I = [-0.5, 0], then a has negative feelings for b. For Potts, the conventionally implicated content of a pejorative is a function that alters the expressive index of a conversational context. So, e.g., if Frank calls George a ‘spic’, the expressive index might shift from , where Frank is indifferent to George, to , where Frank has negative feelings toward George. The lowering of an expressive index could have certain propositional implications (e.g., in the case of Frank and George, we can infer that

Frank is upset with George, that he dislikes Hispanic people, etc.). Potts could argue that a feeling of sullying results from taking part in a conversation whose expressive index has been lowered due to the use of a slur, and negating the utterance would not serve to alter the expressive index, which is not itself a kind of propositional content.

One problem with Potts’s theory is that expressive indices are supposed to measure psychological states of conversation participants, and these can depend on a variety of idiosyncratic features of the participants – their background beliefs, values, etc. This makes it difficult to see how the expressive content of pejoratives could be objective and speaker- independent (Hom, 2010, p. 180). Additionally, Potts’s numerical modeling of attitudes seems too coarse-grained to explain the derogatory nature of different slurs. One could shift the expressive index of a conversation by using non-pejorative terms. Frank might lower the expressive index in a conversation about his colleague, George, by pointing out that George is late to work and that he’s not dressed appropriately for the office. Frank could also lower the index by uttering, ‘Here comes George!’ in a contemptuous tone of voice. Even if expressive indices can model negative attitudes directed at a specific group, it is not clear how such indices could model negative attitudes based on their group-membership. One could be angry with white people, e.g., because one thinks they hold a wrongful social position of power and are afforded certain privileges that are systematically denied to others, not because of their race. If the pejorative content of racial slurs should be understood in terms of expressive indices, Potts will have difficulty explaining the distinctively racist nature of these words.

Finally, it is not clear that expressive indices do any important explanatory work. If Hom

(2010, p. 180) is correct in suggesting that it is the job of traditional semantics, pragmatics, and psychology to account for how the use of expressive language leads to an emotional change in discourse participants, then once this is done, Potts’s expressive indices will seem explanatorily profligate. At most, the theory of expressive indices provides a unique way of describing what has happened when a pejorative is used during a conversation, but the indices themselves do nothing to illuminate how it happened.

8. Perspectivalism

According to Camp (2013), slurs are rhetorically powerful because they conventionally signal allegiance to a derogatory perspective. Perspectives are “open-ended ways of thinking, feeling, and more generally engaging with the world and certain parts thereof” (Camp, 2013, pp.

335-6). They provide “a lens for interpreting and explaining truth-conditional contents,” and tend to motivate certain feelings as natural (Camp, 2013, p. 335). On this view, a speaker who uses a slur for group G signals, inter alia, “commitment to taking the property g that determines the slur’s extension to be a highly central feature in thinking about Gs” (Camp, 2013, p. 337). He also signals “that he is not ‘of’ or aligned with Gs” and “that Gs are not worthy of respect”

(Camp, 2013, p. 338). Within the derogatory perspective, g is taken to be diagnostic in explaining further properties the target is alleged to have (laziness, stupidity, etc.) (Camp, 2013, pp. 337-8). Perspectivalism purports to explain the characteristic rhetorical effects of slurs:

The automatic nature of semantic understanding in general, along with the fact that perspectives are intuitive cognitive structures only partially under conscious control, means that simply hearing a slur activates an associated perspective in the mind of a linguistically and culturally competent hearer. …[The hearer] now thinks about Gs in general…and indeed anyone affiliated with Gs in the slurs’ light, however little she wants to. (Camp, 2013, p. 343)

It is not clear how we should interpret Camp’s claim that simply hearing a slur activates an associated perspective in listeners. Camp’s view admits of a strong version and a weaker version. On the former, hearing a slur with which one is competent automatically makes one think as the bigot does. On the latter, hearing a slur merely disposes one to think as the bigot does. The weaker version holds that overhearing a slur facilitates getting a derogatory perspective, though it does not guarantee that listeners will get one.

Camp is also unclear on what she takes a signal to be. She may have in mind the

56 definition proposed by Green (2007, p. 26): a “signal” is any behavior that makes certain information publicly observable, and was designed for its ability to do so.44 In that case, on

Camp’s view signaling a perspective p is a matter of performing behavior designed to make manifest one’s allegiance to p. On the present reading, perspectivalism predicts that whenever an agent A signals a derogatory perspective p by means of uttering a slur s, observers who recognize

A’s utterance of s as a signal of p will, as a result, think (or be disposed to think) in accordance with p.

The strong interpretation of Camp’s perspectivalism seems implausible and is not empirically supported – data suggest that competent listeners who are exposed to slurs needn’t think that targets are not worthy of respect, that they are not aligned with targets, and so on.

Carnaghi and Maass (2007) found that priming subjects who identified as heterosexual with anti- gay slurs (such as ‘fairy’ and ‘fag’) automatically elicited negatively valenced associations with homosexuality (in comparison with neutral labels ‘gay’ and ‘homosexual’, which did not). They also found that anti-gay slurs did not negatively affect participants who identified as homosexual, who reacted the same way to slurs as they did to neutral labels. On the present interpretation,

Carnaghi and Maass’s findings are problematic for perspectivalism, which predicts that hearing a slur automatically activates a derogatory perspective in all competent listeners, regardless of their group membership.

Another worry is that the strong interpretation of Camp’s view is inconsistent with our

(presumably justified) moral evaluations of racists. It is not clear how we could justifiably hold racists responsible for their views if becoming a racist is as easy as overhearing a slur with which one has minimal linguistic competency. In general, the strong reading proves far too much. If I

44 Green’s definition is a generalization of the definition of a biological signal proposed by Maynard Smith and Harper (2004, p. 4). 57 overhear someone say that my favorite restaurant serves food ‘not fit for a dog’, I can understand the kind perspective they are putting forward, yet nothing forces me to adopt a similar view (cf.

Richard, 2008, pp. 21-22).

Seemingly, Camp should opt for a weaker formulation of her view: hearing a slur disposes competent listeners to have a derogatory perspective on the target. In that case, she owes an account of how this works. Camp points to the “automatic nature” of semantic understanding, but nothing in her formulation of perspectivalism helps us understand how the semantic processing triggered by hearing a slur makes listeners undergo a cognitive, affective, or experiential change that disposes them to view the target as the bigot does. Camp suggests that slurs are rhetorically powerful because in using a slur, a speaker signals a derogatory perspective.

But it is not clear that the notion of signaling a perspective can do the work she needs. Camp acknowledges that speakers can signal derogatory attitudes with a slur even when they have no intention to do so and are ignorant of the word’s meaning (2013, p. 339). However, pace Camp, we have no reason to think that merely signaling a perspective is sufficient for disposing others to think similarly. Thus, even the weak interpretation is too strong. A libertarian could signal her political perspective by placing a ‘Ron Paul’ bumper sticker on her car, yet this behavior is not likely to automatically dispose observers to adopt a libertarian point of view. Likewise, when

George overtly signals his commitment to the Nazi ideology by performing a “Nazi salute” gesture, and I observe George’s gesture, I am not thereby disposed to adopt a white supremacist perspective. Thus, the fact that slurs signal a derogatory perspective (if it is a fact) cannot be the whole story behind their rhetorical power.

However, behavior that signals an attitude a can in some cases lead to the presence of a in observers. A smile can motivate joy in recipients by way of emotional contagion, which

58 occurs when observations of someone else’s emotional display lead us to mimic their facial expression, and this in turn produces a simultaneous congruent emotional response in us

(Doherty, 1997, pp. 131-2; see also Hatfield, Cacioppo, & Rapson, 1993; Fridlund, 1994). Camp might reply that the effects of slurs should be understood along these lines. Perhaps uttering a slur for a group G is like sneering at the targeted group (cf. Hornsby, 2001). When one observers someone sneer at Gs, one is likely to mimic a sneer and one thereby comes to feel angry at Gs.

However, an account of the harm of slurs that uses natural signals as a model is unlikely to provide succor for perspectivalism, since evolved signals do not seem capable of activating anything like a derogatory perspective, as construed by Camp. A perspective is, to use Camp’s metaphor, a lens that shapes and colors our experience of the world. Merely observing, say, an interlocutor’s angry facial expression and becoming angry by means of emotional contagion is not to adopt any sort of perspective. Thus, we still require a specification of a representational medium by which slurs affect competent listeners.

9. Inferentialism

I will now turn to several semantic theories of slurs. Inferentialism is the thesis that knowing the meaning of a statement is a matter of knowing the conditions under which one is justified in making the statement; and the consequences of accepting it, which include both the inferential powers of the statement and anything that counts as acting on the truth of the statement (Dummett, 1981, p. 453). On this view, one knows the meaning of ‘George’s argument is valid’, e.g., if one knows the criteria for applying ‘valid’ to arguments, and one understands the consequences of such an applications, namely that an argument’s validity provides a basis for accepting its conclusion so long as one accepts its premises.

Dummett (1981) offers an inferentialist account of slurs (see also Tirrell, 1999, and

Brandom, 2000). Dummett posits two inference rules for slurs: an introduction rule and an elimination rule. The introduction rule gives sufficient conditions for applying the slur to someone and the elimination rule specifies what one commits oneself to by doing so. Consider the slur ‘boche’, which was conventionally applied to people of German origin (though it is now outdated):

The condition for applying the term to someone is that he is of German nationality; the consequences of its application are that he is barbarous and more prone to cruelty than other Europeans. We should envisage the connections in both directions as sufficiently tight as to be involved in the very meaning of the word: neither could be severed without altering its meaning (p. 454).

Williamson (2009) formalizes Dummett’s inference rules for ‘boche’ as follows:

Boche introduction: x is a German Therefore, x is a boche

Boche elimination: x is a boche Therefore, x is cruel

Brandom (2000) endorses this inferentialist account of slurs, and notes that there is a sense in which slurs are unsayable for non-prejudiced speakers. On his view, once one uses a term like

‘boche’, one commits oneself to the thought that Germans are cruel because of being German.

The only recourse for non-xenophobic speakers, Brandom concludes, is to refuse to employ the concept, since it embodies an inference one doesn’t endorse.

The inferentialist theory is well suited to explain the descriptive features of slurs as well as expressive autonomy. It can also account for why a slur is derogatory toward an entire group of individuals, even when a speaker intends only to derogate a single person in a particular context with the term.

However, the inferentialist’s treatment of slurs is susceptible to numerous objections.

First, Hornsby (2001) questions whether it is even possible, for every slur, to spell out the consequences to which its users are committed. Further, as Williamson (2009) observes, a speaker might grow up in a community where only the pejorative word for a group is used.

Someone might only know Germans as people who are ‘boche’, and not know them as Germans.

In that case, the speaker could be competent with ‘boche’ (she could know that it is a xenophobic term of abuse) without knowing the word ‘German’. Thus, Williamson concludes, knowing the

‘boche-introduction’ rule is not necessary for competency with the slur. Finally, some approbatives seem to lack neutral counterparts (these include ‘stunner’ and ‘knockout’), and if so, it is difficult to see how the inferentialist view can extend to cover them.

10. Stereotype Semantics

According to Hom (2008), the derogatory content of a slur is constituted by its literal, truth-conditional meaning. Hom makes use of the semantic externalist framework first developed by Putnam (1975) and Kripke (1980). Semantic externalism is the thesis that the meaning of a word is not wholly determined by internal states of the particular speakers who use it, but is instead determined (at least in part) by external social practices of the linguistic community in which the word is used. According to Putnam (1975), one can competently use terms like ‘elm’ and ‘beech’ without understanding the complex biological properties of each kind of tree, as long as one stands in the appropriate causal relation to the social institutions that determine their meaning. Similarly, for Hom, the meaning of a slur is determined by a social institution of racism, which is constituted by a racist ideology (a set of negative beliefs about a target group), and a set of harmful discriminatory practices. Hom offers the following formal schema for the semantic content of slurs:

Ought to be subject to p*1 + … + p*n because of being d*1 + … + d*n all because of being NPC*,

where p*1 + … + p*n are prescriptions for harmful discriminatory treatment derived from a set of racist practices, d*1 + … + d*n are negative properties derived from a racist ideology, and NPC* is the semantic value of the slur’s neutral counterpart (Hom, 2008, p. 431). Hom calls his view

Combinatorial Externalism (CE). On this view, ‘chink’ expresses the following complex, socially constructed property as part of its literal meaning: ought to be subject to higher college admissions standards, excluded from managerial positions…, because of being slanty-eyed, devious…, all because of being Chinese.

Hom (2012) extends CE to cover swears. Consider Hom’s analysis of ‘John fucked

Mary’:

to say that John fucked Mary is to say (something like) that they each ought to be scorned, ought to go to hell, ought to be treated as less desirable (if female), ought to be treated as more desirable (if male), ought to be treated as damaged (if female), …, for being sinful, unchaste, lustful, impure, … because of having sexual intercourse with each other. (Hom, 2012, p. 395)

In speech communities where the religious and misogynist ideologies that support the meaning of the term are rejected and more progressive ideas about sex are accepted, the above prescriptions will no longer be a part of the semantic content of ‘fucked’, and the term will come to have a different semantic content.

According to Hom, one motivation for CE is that it accounts for the common intuition that slurs have empty extensions. A non-racist speaker might say ‘There are no niggers; there are only African Americans’. Given that no one ought to be subjected to discriminatory practices because of their race, CE predicts that all racial slurs have null extensions. Hom’s semantic

62 analysis also accounts for expressive autonomy, since the social institutions that determine the meanings of slurs are independent of the attitudes of particular speakers. Finally, CE accounts for non-derogatory, appropriated uses of slurs by in-group members. For Hom, when a targeted group appropriates a slur, they create a new supporting social institution for the term which imbues the term with a new (non-pejorative) semantic content.

One problem is that CE predicts that (34) and (35) mean the same thing, yet one is derogatory while the other is not:45

(34) Niggers are people who ought to be discriminated against because of being African American.

(35) African Americans who ought to be subject to discrimination because of being African Americans are people who ought to be subject to discrimination because of being African American.

Whereas (34) is offensive and derogatory, (35) is trivially true (albeit odd). Embedded uses of slurs also pose a problem for CE, which entails that embedding should eliminate the derogatory power of slurs.46 According to Hom (2012), derogation requires the actual predication of a slur to a targeted individual. However, a speaker who asserts (36) hasn’t literally assigned negative properties or prescribed negative practices for anyone, yet the utterance is still offensive and derogatory:

(36) If he is a chink, he is contemptible.

45 Whiting (2013) gives a similar objection, though he cites different linguistic data than what I have presented here. 46 However, Hom’s semantic view does appear to account for the behavior of other pejoratives under embedding. Take ‘bastard’, for instance. This term does not seem to scope out of conditionals or negation: ‘If he’s a bastard, do not invite him’; ‘He’s not a bastard’. As Hom (2012) observes, someone who utters these sentences need not be interpreted having expressed a negative attitude toward anyone. 63

As Richard (2008, p. 17) observes, if semantic theories like Hom’s are correct, non-racist speakers should be able to endorse utterances like (36), since they would be unproblematically true, given their false antecedents.

In response, Hom (2012) suggests that “speakers don’t normally use a term t unless they are committed to the extension of t as non-empty,” and so hearers will take a speaker who uses a slur in a conditional sentence like (p. 36) to be conversationally implying, inter alia, that the speaker thinks that Chinese people merit discrimination. Notice, however, that one could explicitly cancel that implicature by following up an utterance of (36) with “not that I mean to imply that there are any actually existing people to whom ‘chink’ applies,” since conversational implicata are cancellable. Yet this would not cancel the derogation accomplished by an utterance of (36).

Bolinger (2015) offers a somewhat similar defense of semantic theories in light of concerns about embedding. On her view, the contrastive choice account, the choice to utter a slur for a Group G rather than a non-offensive alternative signals the speaker’s derogatory attitudes toward G (though Bolinger does not understand the signaling involved here in terms of implicature). Bolinger observers that when reporting on a slurring utterance, speakers may have the option to use an abbreviation (‘the N-word’, ‘the C-word’) or some alternative gloss (‘She used a really bad word for Hispanic people’). On this account, the apparent wide-scoping behavior of slurs under embedding is due to the fact that the speaker failed to use an available non-offensive alternative. Thus, the wide-scoping phenomenon applies only in cases where speaker’s choice to use a slur was not forced (other cases where speakers are forced to directly quote someone’s use of a slur tend not to generate offense). As Bolinger points out, this explanation is consistent with semantic theories, and so the behavior of embedded slurs need not

64 pose a challenge to these accounts.

It is not clear, however, that the choice to utter a slur reliably signals anything about the speaker’s attitudes. Take well-intentioned satirical uses of slurs, where a speaker utters a slur in an effort to lampoon and criticize racism. Some such uses may be successful and inoffensive, yet there may be a range of borderline cases about which we do not have clear intuitions concerning the slur’s derogatoriness, even when we have reason to believe that the speaker herself is not prejudiced. Certain pedagogical uses may also be problematic for the contrastive choice account.

The question of whether slurs uttered in “academic” settings are derogatory is a contested, murky issue. I have encountered colleagues who think that speakers should avoid even mentioning slurs whenever possible, whereas others do not find such stringent restrictions necessary. Seemingly, mentions of slurs and other embedded uses can exhibit “leakiness” even when they are not a signal of prejudice in the speaker.

A final worry about CE is due to Jeshion (2013a). She objects that CE’s account of the semantic content of slurs has it backwards – ideologies and social practices must antedate slurs, and this is a problem because the use of a slur for a particular group often plays a role in the creation and development of such institutions and practices. If so, a social institution could not be the source of a slur’s pejorative content.

Hom and May (2013) develop a semantic theory of pejoratives that is somewhat different compared to Hom’s earlier view. They understand the semantic content of slurs in terms of the pejorative concept PEJ, which “denotes a second-level function that combines with a first-level concept (e.g. of race, gender, religion or class) to form a complex first level pejorative concept”

(Hom & May, 2013, p. 298). Sennet and Copp (2015) gloss this view as follows: “for a non- pejorative predicate ‘N’, PEJ(N) is the concept of being an appropriate target of negative moral

65 evaluation on account of being N” (p. 1084). So, for Hom and May, the semantic content of

‘kike’ is given by PEJ(Jew), which generates as its output the pejorative concept, ought to be subject to negative moral evaluation because of being Jewish.

Sennet and Copp (2015, p. 1086) observe that Hom and May’s semantic view does not seem to extend to pejoratives such as ‘dwarf’ and ‘retard’, which do not appear to involve any kind of moral evaluation of their targets. We can add ‘ching chong’ to this list, which is doubly problematic for Hom & May – not only does this slur not suggest any sort of moral evaluation for its target (East Asian people), it does not appear to semantically encode anything.47

Hom and May’s approach avoids Jeshion’s worry that slurs may antedate social institutions of racism (in fact, their view makes no mention of social institutions of any kind), yet it is susceptible to objections similar to those raised against Hom (2008). First, there are problems arising from substitutions of slurs with their neutral counterparts, as in (37) and (38):

(37) No chink will ever teach at this university.

(38) No-one will ever teach at this university who is of Chinese ancestry and who deserves negative moral evaluation simply because he is of Chinese ancestry.

As Sennet and Copp observe, if Hom and May are right, (37) and (38) are equivalent, yet (37) seems derogatory, yet (38) is a bland and non-pejorative claim (Sennet & Copp 2015, p. 1102).

Sennet and Copp also provide the following counterexample (2015, p. 1099):

(39) Kikes deserve negative moral evaluation, but not because they are Jewish (It’s because they are greedy).

47 As I explain in Chapter 4, examples like this also pose a challenge to pragmatic accounts that rely on slurs encoding truth-conditional content that is separate from what they pragmatically express. 66

Hom and May’s theory predicts that (39) will be incoherent, yet it is not (Sennet & Copp 2015, p. 1099).

Certain grammatical construction with slurs generate problems for Hom and May. In the film Malcolm X (Spike Lee, 1992), a character named Rudy asserts the following:

(40) I’m half-wop, I’m half-nigger; I’m not afraid of nobody.

It would be odd to interpret Rudy as having literally said that half of him ought to be subject to negative evaluation because of being Italian and the other half ought to be subject to negative evaluation for being Black, yet Hom and May seem committed to glossing Rudy’s statement this way.

Ironically, one of the explanatory desiderata for a theory of slurs posited by Hom (2008) appears to pose a problem for both Hom (2008) and Hom and May (2013). As Anderson and

Lepore (2013) observe, it is difficult to see how Hom (2008) can explain variation in offensiveness among co-referential slurs. Consider, e.g., ‘guinea’, which Italians consider to be highly derogatory, and ‘dago’, which is comparatively less offensive. The problem is that for

Hom (2008), variation in offensiveness among different slurs was supposed to be explained by differences in the social institutions of racism that support their meaning. Yet Hom and May are committed to holding that both ‘guinea’ and ‘dago’ have the same meaning (namely, ought to be subject to negative moral evaluation because of being Italian).

11. Prohibitionism

Anderson and Lepore (2013a, 2013b) deny that the characteristic features of slurs are due to the contents they express. Their proposal is simply that “slurs are prohibited words; as such, their uses are offensive to whomever these prohibitions matter” (2013a, p. 21). As we have seen,

Anderson and Lepore observe that quotation and various forms of embedding do not always eliminate the offensiveness of slurs, yet content theories seem to predict the opposite. (41), e.g., seems offensive despite the quotational use of the slur it contains:

(41) ‘Nigger’ is a term for Blacks.

They argue that content theorists will have difficulty accounting for the widespread practice of avoiding the word ‘nigger’ completely (using the locution the N-word in place of quoting the term).

While Prohibitionism accounts for the behavior of embedded slurs, it faces several objections. First, Prohibitionism does little to account for the derogatory power of slurs (Jeshion

2013a; Croom 2011). Pointing out that slurs are prohibited words does not help us understand how they are such effective vehicles for spreading prejudice. Anderson and Lepore (2013b) suggest that “[t]he dominant group’s use of the expression might be a vivid reminder of the relation of oppression in which the subordinated group is situated” (p. 7). Yet this is not sufficient for derogation – many things can remind of a history of oppression (take a history textbook, for example).

Anderson and Lepore also suggest that slurs are offensive and derogatory because targets object to being referred to with them (and so it is not just any prohibition on a term that makes it derogatory – the relevant prohibitions come from the target group). On this suggestion, slurs come into existence when a group objects to the use of a name that they did not adopt for themselves. However, as Whiting (2013, pp. 368-369) notes, it is possible for there to be slurs in the absence of taboos or social prohibitions. We can imagine a society in which the vast majority of speakers are prejudiced toward a particular group, and the targeted group members have internalized a racist perspective on themselves, and so they do not object to the use of slurs. We

68 can also imagine that the target group is completely unaware of the use of the slur because bigoted speakers never use it when targets are around. In this case, no one objects to the use of slurs or finds them offensive, yet slurs might still be derogatory. Thus, social prohibitions cannot be all there is to the derogatoriness of slurs. A related worry arises from the possibility of slurs used in foro intero by bigots. We can imagine a racist who creates a slur and uses it only internally, never uttering the term out loud. Under such circumstances, prohibitions could never arise.

Another problem is that Anderson and Lepore’s view rules out a priori the possibility of slurs that should not be prohibited, yet we can imagine instances in which a group should be slurred. As I noted above, ‘fascist’ may be one acceptable slur. We can also imagine a vulnerable

Indigenous group using slurs for people who have colonized their land. Such uses may play an important role in establishing an indigenous group identity, and it is not obvious that such slurs ought to be prohibited.48

A fourth problem with Anderson and Lepore’s deflationary approach is that it fails to generalize to other kinds of charged language, such as approbatives, pejoratives like ‘jerk’ and

‘bastard’, which are often used as playful insults, and borderline cases of jocular slurs that are not widely prohibited, e.g., ‘yuppie’ and ‘ivory tower dweller’ (for academics).

Finally, Anderson and Lepore’s explanation of appropriation is inadequate. They acknowledge that a targeted individual may use slurs for his own group “without violating its prohibition because his membership provides a defeasible escape clause; most prohibitions invariably include such clauses” (p. 20, emphasis in original). But even if Prohibitionists can allow for appropriated uses, it is not clear that they can give a satisfactory explanation of the psychological motivation behind appropriation. If slurs are merely prohibited words, why would

48 I return to this issue in Chapter 6. 69 anyone be interested in appropriating them (as opposed to simply eliminating their use)? The successful appropriation of an epithet takes away a prevalent means by which prejudice against one’s group is propagated, and transforms it into a vehicle for more benign communicative acts.49 Thus, the real motive behind appropriation is eliminating negative attitudes about the target group, and to explain this we may need to appeal to (among other things) some sort of derogatory content that the term promulgates.

12. Conclusion

We have seen that each theory of charged language considered above failed to meet at least one of the desiderata discussed in Chapter 1 (see Table 1 below). All of these views fail to account for the derogatory power of charged language. Part of the problem is that they rely too heavily on positing a single kind of content or mechanism by which slurs denigrate targets. As a result, all of the theories also fail to account for the heterogeneity of slurs.

49 I sketch an account of appropriation in Chapter 3. 70

Table 1:

Derogatory Truth- Expressive Embedded Appropriation Heterogeneity power conditional autonomy uses content Hybrid N Y Y Y Possibly N Expressivism A gestural N Y Y Y Possibly N Theory Truth-value Possibly N/A50 Y Y Possibly N gaps Presupposition N Possibly Y N Possibly N

Fregean N Y N Y Possibly N Minimalism

Conventional N Y Y Y Possibly N Implicature Perspectivalism N Possibly Y Y Possibly N

Inferentialism N Y Y Y Possibly N

Combinatorial N Y Y N Y N Externalism Hom and N Y Y N Possibly N May’s semantic theory Prohibitionism N Y N/A51 Y N N

50 Richard takes himself to explain away apparent truth-conditional content of slurs. 51 Prohibitionists deny that slurs express derogatory content. 71

Chapter 3: Mental Imagery and the Rhetorical Power of Slurs

1. Introduction

Why are terms like ‘slanty-eyed’, ‘wetback’, and ‘redskin’ effective vehicles for promulgating prejudice toward East Asians, Hispanics, and Native Americans, respectively?

What accounts for the effectiveness of figurative slurs, such as referring to garlic as ‘Italian perfume’, and slurs that purport to imitate their targets, such as ‘ching chong’? Embodied cognition theorists argue that interpreting natural language utterances characteristically leads hearers to form mental images or simulate certain experiential states. Drawing on this research, I argue that many paradigmatic slurs and approbative terms are rhetorically powerful because of imagery they invite hearers to entertain.52 Empirical data suggest that imagery has the power to influence our thoughts, emotions, and behavior (Harris, 2000; Gendler, 2006, 2008). Forming a mental image of a ‘slanty-eyed’ East Asian person can lead us to view the target group as exotic and strange-looking.

This is not to suggest that a term must invite imagery in order to be a slur. As I argued in

Chapter 2, theories that posit a single type of derogatory content for all slurs are inadequate because they fail to account for slurs’ diversity. We need not treat slurs as a uniform, homogeneous class. While an embodied approach extends to a large portion of this class, my

52 Invitations (as understood here) needn’t be deliberate acts. Section 1 contains a more detailed discussion of the notion of an invitation used here.

72 view is pluralist in that I argue that no one specific mechanism accounts for the power of all slurs. As I explain below, different slurs may be vehicles for different kinds of derogation.

2. Slurs and Embodied Language Processing

Empirical research indicates that language is an effective tool for prompting imaginative activities of various kinds (see, e.g., Lakoff & Johnson, 1980; Barsalou, 1999; Barsalou &

Wiemer-Hastings, 2005; Lakoff, 2012). The words on the pages of a novel invite readers to imagine the characters and events depicted therein, and advertisements often contain imagery- evoking language designed to influence consumers (Stern, 1988). As we will see, many paradigmatic slurs work similarly. Invitations needn’t be deliberate or effortful. I may invite someone to engage in social interaction by performing a “social” smile, yet such emotional displays are often spontaneous (though I can certainly be held accountable for such behavior).

Non-human primates invite conspecifics to play with them or to participate in grooming with a

“play face” (Green, 2007, p. 119). Such invitations needn’t be performed with a Gricean communicative intention (i.e., an intention to produce an effect in an audience by means of the recognition of one’s intention) in order to be effective. In general, when we invite others to imagine something, we needn’t take ourselves to be introducing an imaginative activity or even be able to articulate what we have invited others to imagine (Walton, 1990, p. 38). A child who hands an interlocutor a doll and asks them to “feed” it may not think of himself as inviting the interlocutor to participate in a shared make-believe, yet this is what the child is doing.

Analogously, invitations to imagine by means of using slurs may be issued spontaneously and non-deliberately.53

Some brief remarks on imagining will be useful. Following Van Leeuwen (2013), I

53 As I explain in Chapter 5, this is also true of invitations to imagine issued by means of derogatory gestures. 73 distinguish propositional imagining, i.e., the process of taking a cognitive attitude other than belief toward a proposition p that treats p as somehow fictional (e.g., treating the proposition that

Harry Potter went to Hogwarts as fictionally true), from constructive imagining, i.e., the process of forming a mental representation that may include imagery and other experiential states.

Constructive imagining does not entail propositional imagining. A detective could attempt to solve a crime by imaginatively simulating a scenario that depicts various ways in which a suspect may have committed the crime in question, and this does not require taking an attitude toward a proposition (Van Leeuwen, 2013, p. 221). We can form imagery in just about any sensory modality (including the visual, auditory, tactile, olfactory, and gustatory modalities), though we may imaginatively simulate certain experiences, e.g., a feeling of weightlessness, without using any sensory modality (Gendler, 2011). As John Locke and Thomas Reid observed long ago, an image may be a static depiction, like a snapshot of an event, or a dynamic mental representation (Barsalou, 1999, p. 581). We can form a dynamic of image of someone swinging a bat or diving into a pool, for instance.

The derogation accomplished by figurative slurs seems to rely, in part, on imagery they evoke in hearers.54 Examples include ‘tar baby’, ‘gatorbait’ (a slur for African Americans commonly used in the southern U.S.), ‘tree hugger’, ‘fairy’, ‘trailer trash’, ‘snot nose’, and

‘Jewish American Princess’, among others. I will not offer a theory of figurative language here, though I will adopt some widely accepted assumptions about the rhetorical power of metaphor. It is uncontroversial that novel metaphors have the striking power to influence us. In 2004, the animal rights group PETA drew a great deal of ire when they published an ad featuring pictures

54 Camp (forthcoming) develops an account of metaphorical insults by extending her perspectival account developed in her (2013), though she does not discuss figurative or metaphorical slurs. Many of the objections raised against Camp (2013) will also apply to her more recent account of metaphor. I discuss Camp’s account and show how my own view diverges from it in more detail in Section 4 below. 74 of emaciated animals in a slaughter house accompanied by the phrase ‘the Holocaust on your plate’ (Shachar, 2014). Novel (non-conventionalized) metaphors like this one evoke imagery that can have deep cognitive and emotional effects on us (Moran, 1989).

Sperber and Wilson (1986) offer a cognitive theory of metaphor based on their relevance theory. On their view, interpreting a metaphor is a matter of forming an ad hoc concept, which helps the hearer get clear on what the speaker meant. As Carston (2010) observes, this approach may work for many outdated, conventionalized metaphors, yet it does not provide an exhaustive theory, since extended literary metaphors seem to require imagery in addition to an adjusted concept (I discuss the distinction between novel and conventionalized metaphors in more detail below). Cognitive theorists may claim that imagery itself is a cognitive phenomenon in that it is reducible to propositional or conceptual content. McGinn raises compelling objections to this move. First, images have a qualitative component: “There is something that it is like to have a visual image, and this something is in important respects similar to what it is like to have a visual percept: both, after all, are aptly described as visual. Both involve simultaneous presentations of a number of visually detectable features – color shape, and so on” (McGinn, 2004, p. 37).

However, the kind of mental state one is in when merely considering propositions about an object is different from forming a visual image of it; in general, considering propositions about an object does not require imagistically representing it (McGinn, 2004, p. 37). Additionally, if

Wittgenstein was right in thinking that one “cannot simultaneously have a percept and an image whose content is the same,” then it seems that imagery is not propositional (McGinn, 2004, p.

37). The idea is that if one is seeing a red cube, one cannot simultaneously form a qualitatively identical image of that very same red cube, yet one can entertain various propositions about, and apply concepts to, a red cube one is seeing. Since propositional and conceptual content “can

75 combine with the percept, but the image cannot,” images are not reducible to propositional or conceptual content (McGinn, 2004, p. 38).55

Because the interpretive processes that lead us to form the relevant imagery can occur quickly and spontaneously, metaphors often move us in ways that elude rational scrutiny

(though, as I argue below, they needn’t do so). Locke (1995) may have sensed this when he labeled metaphors “perfect cheats.” As we saw in Chapter 2, Gottlob Frege argued that the rhetorical effects of metaphors and other forms of poetic language are best understood as part of their “coloring,” which is constituted by subjective ideas, feelings, and images that are distinct from a term’s sense and reference. Frege was primarily interested in developing a rigorous formal language that would illuminate logical relations among statements. Thus, part of his motivation for characterizing derogatory content as subjective coloring was a desire to cordon off emotionally charged terms from what he took to be the proper subject matter of philosophy of language. Like Frege, Davidson (1978) holds that metaphors often have an imagistic element that is non-semantic. The account I shall develop is broadly Fregean and Davidsonian in that, for a slur whose rhetorical power relies on imagery, I take it that the imagery is not, in any sense, part of the slur’s conventional meaning. Further, I am not suggesting that slurs semantically encode an invitation to form imagery (such that a slur for group G literally means something along the lines of ‘hearers are invited to imagine such-and-such about G’). Rather, inviting imagery is something a speaker does with a slur. Of course, we need not hold, as Davidson did, that the imagistic effects of metaphor are like a “bump on the head” (1978, p. 46). As contemporary Davidsonians acknowledge, a metaphor’s rhetorical effects often depend on

55 Moreover, if embodied cognition theorists are correct, conceptual processing itself often involves imagery and perceptual simulation. Empirical studies suggest that conceptualizing actions, motion, and object form activates brain centers for perceiving form, centers for perceiving motion, and the motor system, respectively (Barsalou et al. 2003, p. 87). However, it does not follow that concepts are identical to images (I discuss additional empirical data on the relationship between images and concepts below). 76 hearers exploring metaphorical imagery, yet they also hold that this needn’t depend on listeners inferring any special propositional speaker meaning from what the speaker literally says (Lepore

& Stone, 2010, p. 170). Moreover, the audience’s participation needn’t be deliberate or effortful.

As Walton (1990, p. 14) observes, when imagining is deliberate, we are well aware of its dependence on us; however, imagining often happens spontaneously, e.g., when we effortlessly start daydreaming about being wealthy, and call up an image of a large mansion and expensive cars. Walton notes that in such cases, we are more spectators than perpetrators of our imaginings.

Theorists standardly distinguish so-called “live” metaphors from “dead” metaphors.

Davidson (1978, p. 38) suggests ‘he was burned up’ as an example of the latter. As metaphors become conventionalized over time, their capacity to evoke vivid mental imagery diminishes. In contrast, it is difficult to read the sentence, “depression was a toad that squatted wetly on your head until it finally gathered the energy to slither off” without forming a vivid image (Carston,

2010, p. 307).56 This is not to suggest that metaphors must be either “live” or “dead”; perhaps a better way to metaphorically delineate the different types of metaphor would be to say that some are fresh and highly imagistic, while others are more-or-less stale and conventionalized, and it is not hard to find intermediate cases (the metaphor of storing data on a “cloud” may be one example). As Lepore and Stone observe, this need not pose a challenge to theories of metaphor such as Davidson’s, since the fact that a metaphor can take on a new propositional meaning does not entail that it always had some special propositional meaning distinct from its literal meaning

(2010, p. 178). Even after a metaphor takes on a new propositional meaning, residual imagery

56 Although she agrees with Davidsonians that novel metaphors are rich with imagery, Carston (2010) maintains that metaphors carry special propositional meanings that are distinct from their linguistically encoded meaning. Consider: “The fog comes/on little cat feet,” due to Carl Sandberg. According to Carston, “on an initial reading, the literal meaning takes over and is metarepresented…along with the mental image of a large, soft, lightly padding but purposefully moving cat, and from these together are derived implications about the way the fog looks and feels” (2010, p. 310). 77 can remain, and Davidsonians distinguish the propositional meaning that stale, conventionalized metaphors take on from the more-or-less vivid images they evoke.

Allen (1983) gives several examples of metaphorical slurs (though he does not offer an account of their rhetorical power). Imagine a speaker denigrating Italians with the metaphor of garlic as “Italian perfume” (Allen, 1983, p. 11). Getting the full rhetorical effects of this metaphor requires, inter alia, imaginatively simulating the experience of getting a waft of a garlic scent emanating from an Italian person. A wide range of images and experiential states in different modalities may be sufficient, yet there are some restrictions that determine what counts as appropriate imagery. In the case of garlic as “Italian perfume,” objective facts about what garlic smells like (and more specifically, what it smells like on someone’s breath or on their hands) provide one constraint. If one imaginatively simulates a pleasant floral scent in response to the metaphor, one has failed to comply. In general, when hearers imaginatively respond to language, they draw on memories derived from perceptual experiences (Barsalou, 1999, p. 592).

In the present case, the relevant perceptual experience may be of smelling garlicky breath or smelling garlic on someone’s hands after they have been cooking. Hearers may also need to draw on background assumptions and stereotypes, e.g., that Italian cuisine is particularly rich in garlic, when imaginatively responding to the slur in question. Thus, even though imagery itself is a subjective phenomenon, metaphorical imagery can be more or less standardized.57

A somewhat more conventionalized example of a metaphorical slur is ‘Jewish American

Princess’ (hereafter, abbreviated as ‘J.A.P.’). Responding to ‘J.A.P.’ by forming a mental image

57 Here I am suggesting only that stereotypes may place constraints on the kinds of imagery that slurs invite, not that all slurs rely on stereotypes. Blum (2002, p. 252) distinguishes stereotypes that attribute a specific set of characteristics to group from stereotypical images that depict them in an unflattering and insulting way, yet needn’t attribute a set of properties to them. Saka (2007) and Jeshion (2013a) observe that a slur’s target group needn’t be associated with any stereotype whatsoever. As I explain below, listeners needn’t always rely on stereotypical properties when imaginatively responding to slurs. 78 of a person that one takes to be a typical young Jewish-American woman receiving lavish gifts from her parents and throwing tantrums when she fails to get her way would be sufficient for complying with an invitation to form metaphorical imagery (I am not suggesting that these features are typical of actual Jewish women). Specific imagery may include a visual image that depicts the target wearing expensive clothing, an auditory image of the target speaking in a whiny voice, making unreasonable demands, and so on. Listeners could use ostensible targets as props when imaginatively responding to slurs, though the use of an ostensible target is not mandated. If we responded to ‘J.A.P.’ with mere propositional imagining, e.g., considering the proposition that Jewish-American women are rich, spoiled, and materialistic as fictionally true and not forming any mental image whatsoever, we would miss the metaphor’s rhetorical point.

As we will see (in Section 4 below), forming unflattering mental images of the target group facilitates thinking of them in a derogatory way, which in turn disposes us to treat them accordingly.

Some slurs are extended metaphors. ‘Ape’-metaphors are commonly used to dehumanize

African Americans (Goff et al., 2008). A speaker could create an extended metaphor by calling

African Americans ‘knuckle-dragging, chest beating apes’. Extended metaphorical slurs do not merely invite certain images of the target; they invite us to savor those images and fill in details.

The degree to which a listener feels sullied, and the degree to which the slur is effective in transmitting prejudice to the listener, may be due to the extent to which she follows the racist speaker down the garden path. This is not to suggest that, in general, imaginative responses to slurs must be deliberate. Given the prevalence of implicit biases against the targets of commonly used slurs,58 and the fact that imagining can be non-deliberate, listeners may find themselves

58 Kelly and Roedder (2008) give an overview of studies on implicit racial bias. 79 imaginatively responding to a slur quickly and effortlessly.59 However, as a metaphorical slur becomes antiquated, it may lose its associated imagery, and consequently, its offensiveness.

‘Limey’, e.g., is a conventionalized metaphor, and as a result its capacity to evoke an image of an

English sailor with scurvy (if it ever invited such an image) is now greatly diminished.

Imagery also accounts for the rhetorical power of what I call descriptive slurs, whose literal meanings provide unflattering descriptions of their targets. ‘Slanty-eyed’, ‘raghead’, and

‘redskin’, invite an unflattering visual image of the target’s eyes, attire, and skin, respectively.

Descriptive slurs may also invite olfactory and auditory imagery. Consider ‘curry muncher’, a slur for various South Asian groups commonly used in the UK. This slur is likely to be influential for listeners who form a vivid image of chewing and slurping sounds, a strong odor emanating from the target’s mouth, and so on. As with metaphorical slurs, the details of the imagery are to some extent up to hearers. Listeners are not told exactly how the target’s eyes, skin, and attire are supposed to look, how they are supposed to smell, and so forth; in order to comply with the speaker’s invitation, they must generate the image themselves. Since the perceptual systems of different hearers may be subtly different, token images may exhibit certain idiosyncrasies. Barsalou (1999) notes that while all people with normal color vision have the same perceptual system, each simulator is implemented in a somewhat different way. This will lead to subtly different simulations. When two people are searching for bananas in a grocery store, e.g., each can simulate the particular idiosyncratic experience of yellow that they are likely to perceive when encountering the desired object (Barsalou, 1999, p. 599). As with numerous metaphorical slurs, complying with an invitation to imagine in response to descriptive slurs may require drawing on relevant stereotypes. In this way, slurs may serve to reinforce latent prejudice

59 Camp (forthcoming) suggests that metaphors amount to an “invitation to intimacy” between speaker and interlocutor. She follows Moran (1989, pp. 90-1), who holds that metaphors affect us “at a level beneath that of deliberation or volition.” I argue below that this is not an essential feature of metaphor. 80 in hearers (though, pace Camp, interpreting slurs does not require adopting a derogatory perspective on the target).

A third category of slurs are not figurative, do not rely on any truth-conditional content, and instead denigrate by mimicking alleged features of the target. I shall call these imitative slurs. ‘Ching chong’, e.g., purports to be iconic of the way that Chinese people sound.60 By purporting to mimic Chinese speakers with this nonsensical phrase, the speaker invites listeners to entertain an unflattering auditory image of the target, and doing so may lead them to hear actual Chinese utterances as silly or strange-sounding. ‘Ching chong’ is a paradigm example of a slur whose rhetorical power cannot plausibly be captured in terms of any semantically encoded content. Further, Grice’s notion of “speaker meaning” does not elucidate this slur’s derogatory power; in using this phrase to denigrate Chinese people, the speaker need not mean that Chinese languages sound strange. In general, the derogation accomplished by stylized imitations eludes verbalization.61

Research on embodied cognition provides an empirical basis for an account of slurs’ rhetorical power that appeals to imagery. Studies indicate that the interpretation of metaphorical language activates imaginative circuitry that is the same for sensory perception and action

(Lakoff, 2012, pp. 778-9). The metaphor of affection as warmth, e.g., crucially depends on the embodied experience of feeling warm when being held affectionately (Lakoff, 2012, p. 777). A study by Matlock (2004) indicates that interpreting metaphorical sentences about motion, such as

‘The road runs along the coast’, led readers to mentally simulate certain actions and experiential states, such as visual scanning.

60 Kat Chow suggests that ‘ching chong’ is a crude imitation of what English speakers think Mandarin or Cantonese sounds like in a post for the National Public Radio blog, Code Switch, accessible at . 61 Chapter 5 contains an account of the rhetorical power of derogatory imitations. 81

Interpreting sentences that describe an observable event also characteristically leads to the activation of imagery and perceptual simulation. In Zwaan et al. (2002), subjects read a sentence, e.g., ‘There was an eagle in the sky’, and were then shown an image, e.g., of an eagle with wings spread, as though it were in flight, or a picture of an eagle with its wings folded, as though it were perched. The subjects were asked to judge whether the image they saw depicted something mentioned in the sentence. Zwaan and colleagues report that subjects who saw an eagle with its wings spread were faster to identify it as something mentioned in the sentence than subjects who saw an eagle with folded wings. They concluded that reading the sentences led subjects to form a mental image depicting what the sentence described, and slower responses were due to a mismatch among the mental image subjects formed and the picture they saw. In

Borghi et al. (2004), participants correctly responded faster to yes/no questions about whether a term (e.g., ‘roof’, ‘headlight’) names a part of a car when ‘yes’ responses required moving the hand in the same direction one would move when physically interacting with the part named.

Borghi et al. conclude that reading the words led participants to simulate certain perceptual experiences and activated relevant action schemas (see also Glenberg et al., 2009). In light of the empirical data on embodied language processing, it is plausible that imagery and simulations of various experiential states are triggered by figurative slurs (including extended live metaphorical slurs, more conventionalized metaphors like ‘Jewish American princess’, and everything in between), as well as descriptive and imitative slurs.

We needn’t infer from these data that language understanding is partially constituted by the cognitive processes that underlie action, sensory imagery, perception, and various other experiential states. Some critics object that embodied cognition theorists have failed to distinguish (i) the notion that mental images and simulations of perceptual states are constitutive

82 of a hearer’s interpretation or understanding of an utterance, from (ii) the notion that interpreting language merely causally contributes to or facilitates the formation of images and perceptual simulations. Critics argue that the empirical data establishes only the latter, yet many embodied cognition theorists purport to have established the former.62 I will not assume that (i) is true, either about cognition generally or about language processing in particular; it would be enough for my purposes if hearing figurative, descriptive, and imitative slurs merely facilitated forming imagery and simulating various experiential states.

It is worth considering whether an embodied language approach could be extended to cover slurs other than the central cases mentioned above. Sometimes a sign or symbol owes its communicative effects to ritualization, which is the evolution of a cue (i.e., a behavior that happens to convey certain information) into a signal (i.e., a cue designed to provoke a certain response in observers) (Maynard Smith & Harper, 2004). In an example of ontogenetic ritualization described by Tomasello (2008), an infant chimpanzee who wants to be carried by her mother will grab onto the mother’s shoulders and pull down in order to climb on. Once mothers come to anticipate what the infant wants, they begin to lower their backs when the initial part of the sequence is produced – i.e., when the infant simply touches the mother’s shoulder – and once the infant learns to anticipate her mother’s anticipation of what she wants, she may use the shoulder touch to signal a request for a ride (Tomasello, 2008, p. 25). A process analogous to ritualization may have imbued ‘nigger’ with the power to invite derogatory imagery, though I will not settle this issue here. There is a historical relationship between ‘nigger’ and racist caricatures of African Americans. From the mid-nineteenth until the mid-twentieth century, the term commonly appeared in print ads, brand logos, cartoon strips and Blackface minstrel shows

62 Adams and Campbell (1999, p. 610), e.g., raise an objection like this to the embodied theory of conceptual knowledge proposed by Barsalou (1999). 83 that depicted African Americans in unflattering ways, e.g., having caricatured facial features, being dim-witted, poor, or even subhuman (Lott, 1993). Such depictions are illustrated below in figures 1 and 2: 63

Figure 1: Figure 2:

Notice the clown-like features in the advertisement in Figure 1, and the enlarged facial features in Figure 2. Large bright red lips were characteristic of the makeup employed by actors who performed in “Blackface” minstrel shows from roughly the mid-nineteenth until the early- twentieth centuries (Lott, 1993). Moreover, in the image on the label in Figure 2, the man is depicted voraciously enjoying an oyster in an almost animalistic way. These portrayals have a legacy in contemporary media where the depictions are sometimes even less flattering (Bucholtz

& Lopez, 2011). Suppose that as a result of viewing minstrel theatre and seeing racist advertisements, listeners became accustomed to seeing a racist image when hearing ‘nigger’, and they began to respond to the term by mentally constructing racist imagery, even when such imagery was not shown to them. Speakers who anticipated this could then use the term to solicit

63 Images in figures 1 and 2 were archived by the Jim Crow Museum of Racist Memorabilia at Ferris State University, and are accessible at .

84 hearers to form the relevant imagery, in a way analogous to how certain instrumental behaviors and practical actions become ritualized into signals.64

For some slurs, associated imagery may come from government propaganda (see Figures

3 and 4 below).65 Notice that in addition to depicting Japanese soldiers as sadistic and cruel (in

Fig. 3), both images include caricatured facial features:

Figure 3: Figure 4:

It is easy to see how frequent exposure to images like those in Figures 3 and 4 could have made

Americans ready to attribute shiftiness to their Japanese neighbors, and to be on the lookout for cruel, threatening behavior by them.

64 The comparison with ritualization here is merely an analogy. A slur being used to prompt recipients to form an image in the way just described would not be an actual instance of ritualization, since this does not involve a behavior’s being segmented. 65 Figures 3 and 4 contain images of posters displayed in the U.S. during World War II. Both were catalogued by the U.S. National Archives and Records Administration. The image in Figure 3 is accessible at: and the image in Figure 4 at . 85

Hom (2008, p. 428) claims that, in general, as targets “integrate back into society and active discrimination subsides,” slurs’ derogatory power often diminishes. If so, it is plausible that once World War II ended, and American exposure to Japanese people and their culture increased, racist depictions such as those in Figures 3 and 4 became less frequent, and the imagery associated with ‘Jap’ faded (though, as I explain below, the slur may still be highly offensive even after this has happened). In some cases, however, outdated imagery gets replaced by new imagery. In the late 1970s the ethnic slur ‘guido’ may have invited something like an image of Tony Manero, the working-class, disco-dancing, Italian-American lead character in the film Saturday Night Fever, and while that image has since faded, the slur may now evoke a very different image due to its use on the reality TV series Jersey Shore (Cohen, 2010). Today

‘guidos’ no longer wear white leisure suits and participate in disco dancing competitions; now they are tanned, tattooed, and wear muscle shirts and gold chains.

Slurs needn’t harm by inviting unflattering images of their targets. Another possibility is that slurs are vehicles for prompting a mental simulation of certain experiential states. Recently, a number of embodied cognition theorists have argued that certain mental phenomena typically thought of as incompatible with embodied theories are in fact grounded in simulations of experiential (or other internal) states. Barsalou and Weimer-Hastings (2005) found that processing abstract concepts often led subjects to report on evaluations, affective states, and desires. The concept truth activated an introspective state, e.g., the subject’s sense that a statement is correct. Processing the concept pigeon activated various evaluative and affective states, such as finding pigeons “dirty” and “gross”, and processing cook, led to descriptions of actions, goals, and desires, e.g., “to make exactly what you want to eat” (Barsalou & Wiemer-

Hastings 2005, p. 159).66 Barsalou and Wiemer-Hastings suggest that one way in which concepts

66 Barsalou and Weimer-Hastings report that the “percentage of introspection properties was higher for 86 could be embodied is that we form memories of various “situational perceptions” about settings, entities, events, introspections, etc., “and the conceptual system…then later simulates these memories to represent concepts” (p. 157). Recent neuroimaging experiments lend some support to this view. Wilson-Mendenhall et al. (2013) report that processing the concept convince activated brain areas associated with inference, social interaction, and certain affective states. We needn’t assume (as Barsalou and Wiemer-Hastings do) that the content of all concepts is partially constituted by embodied simulation. If theorists are right that processing concepts leads to simulations of various kinds (e.g., of certain negative evaluative or affective states), it is not implausible that exposure to slurs could do the same. For so-called ‘untouchables’, the lowest- ranking groups in societies organized by a caste system, perhaps the relevant mental states would be a sense that the target is pitiable or disgusting, or a feeling of aversion (while I take this suggestion to be a plausible empirical hypothesis, establishing it will require further empirical investigation, and so the question of whether ‘untouchables’ prompts an embodied simulation of the attitudes describe above is not the kind of issue we can settle here).

This is not to suggest that imagery of one sort or another is always crucially involved in the derogation accomplished by slurs. Given that slurs make up a diverse class, a pluralistic approach is needed. As I noted in Chapter 1, slurs may be vehicles for a visceral kind of derogation – non-prejudiced listeners may feel shocked or offended when overhearing them. A slur may start out highly imagistic, and lose its imagery over time. One plausible example is

‘coon’, which is traceable to minstrel show portrayals of African Americans from the nineteenth and early twentieth centuries (Lott, 1993). Even after the cultural depictions that supported this

abstract concepts (28%) than for concrete and intermediate concepts (21%, 22%)” (p. 152). They also report that concrete concepts often produced situational representations of various activities, in addition to descriptions of introspective states. Relevant situational information activated by the concept cars included information about driving, fueling, and washing (Barsalou & Weimer-Hastings 2005, p. 156). 87 slur’s imagery are no longer present, and the original basis for finding it offensive is forgotten, it may still damage us in a visceral way by shocking or offending targets and bystanders. People may still find ‘coon’ highly offensive even if they have no knowledge of minstrelsy (though under the right circumstances the slur could regain its imagery and it may evoke certain other experiential states even after the minstrel caricature it once invited is gone). This is an instance of a more general phenomenon that Tomasello (2008) calls the “drift to the arbitrary.”67 Targets and other contemporary listeners who happen to be present when the taboo on the term is violated may feel a kind of visceral discomfort, and targets may feel personally disrespected (cf.

Anderson & Lepore, 2013b), yet it is unlikely that hearers will be sullied by the use of the term.68

A non-imagistic slur’s derogatory power may be due, in part, to its syntactic features.

Lycan (2015, p. 5) notes that many Australians find the use of ‘Abo’ highly offensive. According to Lycan, this term is a diminutive form of ‘Aboriginal’. The fact that targets consider this term a slur may be due, in part, to the history of oppression and institutional racism faced by Aboriginal

Australians. Until the 1970s, Aboriginal children were removed from their families and put into foster care by the Australian government (Killmister, 2012, p. 256). As Anderson and Lepore

(2013b) observe, the offensiveness of a slur sometimes depends on who introduced it. We should expect oppressed groups to object when oppressors refer to them in a way usually reserved for friends, and in the present case, this occurs when oppressors use the diminutive form of

67 Tomasello focuses on a drift to the arbitrary that occurs with gestures. As he notes, iconic gestures tend to become arbitrary communicative vehicles as they become conventionalized over time. I discuss this phenomenon as it applies to derogatory gestures in Chapter 5. 68 However, pace Anderson and Lepore, in general, a target group G’s objecting to the use of a term t, which refers to Gs, is not a necessary condition on t’s being a slur for Gs. As I noted in Chapter 2, we can imagine a racial group who has internalized racist attitudes for their own group, and so they view the use of slurs that target their group as entirely appropriate. Or the target group may simply not know about the slur. In both cases, the slur in question would not be taboo, yet it may still be derogatory. 88

‘Aboriginal’ to refer to Aboriginals.69 Also consider ‘Paki’, seemingly an abbreviation of

‘Pakistani’. This slur emerged in Great Britain the mid-twentieth century while Pakistan was in the midst of a struggle for independence from British rule, and is still considered highly offensive by Pakistani immigrants in the U.K., who face various forms of discrimination even today (Allen, 1983; Hughes, 2006). It is not difficult to see why Pakistani immigrants would object to the use of ‘Paki’ by members of an oppressive majority group. In comparison, given how long relations between the U.S. and U.K. have been friendly, it is not surprising that the use of the epithet ‘Brit’ (an abbreviation of ‘British’) by Americans is unlikely to provoke offense.70

This is not to deny that ‘Abo’ and ‘Paki’ could also be vehicles for prompting an embodied simulation of negative affective state directed at Aboriginals or Pakistani people or other kinds of imagery that depict these groups in an unflattering way. The point is simply that these terms are at least vehicles for the visceral kind of derogation described in Chapter 1 – people find the use of these terms shocking, offensive, and insulting because targets object to their use. In comparison, figurative, descriptive, and imitative slurs are more effective vehicles for sullying hearers, though they may also facilitate the more visceral harm described earlier. If all slurs were vehicles for the latter sort of derogation, then an appeal to our imaginative capacities in an explanation of their derogatory power would not be justified. However, in light of the empirical data discussed above, it is plausible that the derogatory power of some slurs crucially relies on imagery.

69 Anderson and Lepore’s distinction of acts of slurring from conventional slurring words (explicated in Chapter 1) is important here. When a white Australian addresses an Aboriginal as ‘Abo’, this may constitute an act of slurring, yet it does not following that ‘Abo’ is a conventional slurring word. However, even if ‘Abo’ is not a conventional slur, it could become one. 70 I am assuming that, in general, people to whom ‘Brit’ refers, namely citizens of the United Kingdom, do not find the term offensive and do not consider it a slur. I know of no empirical studies that support this assumption, however the online version of the Oxford English Dictionary defines ‘Brit’ as “A British person,” and classifies the term as a colloquialism, rather than a pejorative: . 89

3. Other Pejoratives and Approbatives

Like figurative, descriptive, and imitative slurs, many other pejoratives and approbatives invite imagery that plays a crucial role in the derogation or approbation they accomplish.71

Rather than offer an exhaustive account of the power of these categories of charged language, I will show how my approach accounts for paradigm examples of each category (though what I say about the paradigm cases also applies to many other members of each category).

Some pejoratives function to denigrate people on the basis of their personal appearance

(Bach, 2014). These include words and phrases that invite an unflattering visual image of a target, such as ‘mouth breather’ and ‘slack-jawed’. Other examples are figurative uses of language. Imagine a speaker denigrating someone with ‘tightass’. The experience of being so tense that one is tightly clenching one’s buttocks gives us a sense of the kind of feature the speaker is attributing to the target. These words and phrases may be more or less conventionalized. In most contemporary uses, the derogation accomplished by ‘asshole’ does not depend on hearers forming an image of an anus. Rather ‘asshole’ is the sort of word one says when one wants to be rude to someone, and so it is mostly a vehicle for the visceral sort denigration (shocking, offending, insulting, and the like) and a less effective vehicle for sullying.

However, one could inject imagery back into this insult, as in:

(1) He’s as unpleasant as an improperly wiped asshole!

Another somewhat stale metaphor is ‘brown noser’, which is customarily used to denigrate someone for being fawning and obsequious. However, as the following example shows, it is not a fully arbitrary insult – one could extend the metaphor in such a way as to revive the “head up the rectum” image:

71 Bach (2014) provides an extensive list of insulting terms and laudatory phrases. 90

(2) He’s such as brown-noser that he has yet to remove his head a week after meeting with his boss.

Also consider the following joke:

(3) The difference between a brown noser and an ass kisser is depth perception.72

Many pejoratives that include ‘fuck’ began as metaphors. In an era in which speakers held that being literally fucked (i.e., having sexual intercourse) made one impure, speakers may have used ‘fuck’ (or ‘fucked’) metaphorically to convey something about a target being worthless, damaged or inferior (Hom, 2012, p. 398). Consider (4)-(6):73

(4) I just saw George’s sales numbers – he’s fucked.

(5) I’m going to fuck you up.

(6) Don’t fuck up the mural you’re painting.

While the connection between ‘fuck’ and sexual violation imagery has become opaque over time,74 insults involving ‘fuck’ may be extended so as to reintroduce past sexual violation imagery (and perhaps also introduce new related imagery). Take (7), for instance:

(7) Go fuck yourself.

While this phrase may look like a stale metaphor, we can imagine a speaker extending the insult by telling her interlocutor to fuck himself with a particular instrument in a way that would be

72 This joke can be found online at . 73 Hom (2012, p. 398) provides additional examples like these. 74 Hom (2012) considers the possibility that statements of the form ‘x is fucked’ are conventionalized metaphors, which literally say that x is damaged. However, rather than arguing that the derogatory content in cases like (4)-(6) is now codified in the term’s literal meaning, Hom suggests that ‘fuck’ has retained its sexual meaning, and glosses (4)-(6) as conversationally implying something to the effect that the target is damaged or inferior. 91 especially painful, so as to invite a vivid image of sexual violation. In that case, mental imagery would account for the utterance’s power to sully listeners.

Pejorative phrases needn’t encode a new literal meaning once they become conventionalized. ‘Go to Hell!’ is not a metaphor – a speaker who denigrates someone with this phrase needn’t be interpreted as comparing the act going to hell to some other action that she is enjoining the recipient to perform. At one point, ‘Go to Hell!’ may have been a vehicle for prompting recipients to imagine being sent to one of the lower circles of the inferno in Dante

Alighieri’s Divine Comedy, or perhaps hearers relied on vivid descriptions of Hell in the Bible or some other holy text. Regardless, in contemporary uses, ‘Go to hell!’ is simply a rude phrase to say to someone. It is not the sort of phrase we would like others to say to us, and so it is a vehicle for a visceral kind of denigration – we feel offended or insulted, as opposed to sullied.

As noted in Chapter 1, approbatives are words whose conventional function is to laud their targets, rather than denigrate them. Such phrases may laud on the basis of personal characteristics, group membership, or both. Approbatives share many features with pejoratives

(see Chapter 1). An approbative phrase may be more or less imagistic/conventionalized, and a drift to the arbitrary is likely to happen with these terms, just as it happens with the pejoratives mentioned above. Consider the (somewhat stale) metaphor of a woman as ‘goddess’.

Descriptions of goddesses in poetry and depictions in paintings, sculptures, and films all provide constraints on the sort of imagery listeners are to construct in response to hearing a speaker predicate ‘goddess’ of someone. Also consider ‘stunner’ and ‘knockout’, whose power to praise depends crucially on an experiential state – the target’s beauty is so intense and striking that one gets an experience analogous to being stunned or briefly knocked unconscious (Emanatian, 1995,

92 p. 173). Because imagery can change over historical time, a term may transition from being a pejorative to an approbative (I return to this issue in Section 5 below).

4. The Power of Imagery

Imagery influences us by means of contagion, which occurs when we respond cognitively, emotionally, or behaviorally to imagined content as if it were reflective of reality

(Gendler, 2006, p. 186). Historically, ‘contagion’ has been used in social psychology literature to describe the transmission of emotions in interpersonal cases. Subjects tend to mimic an interlocutor’s expression of emotion in the face or voice, and as a result they come to feel emotions similar to those expressed by their interlocutor (Hatfield, Cacioppo, & Rapson, 1993).

Gendler describes cases of intrapersonal imaginative contagion that occur when imagining something influences how one thinks, feels, or behaves. We quarantine imagery (and other kinds of imagined content) from our other mental representations when we restrict its effects to the domain of our imaginative exercise, and we are often successful – even children do not believe that spilling imaginary tea will result in a table actually becoming wet (Gendler, 2011; Leslie,

1987). However, we regularly fail to quarantine imagined content because numerous features of our mental architecture are “source indifferent,” in that they “process internally generated and externally generated content in similar ways – even in cases where the content in question is explicitly ‘marked’ as reality insensitive” (Gendler, 2006, pp. 183-4).

Contagion can occur in imagining that is purely content-based (i.e., not guided by external props). Just as staring at a black square can generate an after-image, namely a white square in one’s visual field, “merely forming a vivid mental image of a black square could have the same consequence” (Gendler, 2006, p. 189). In an experiment on bystander apathy, Garcia et al. (2002) found that subjects instructed to imagine being part of a group of ten people offered

93 less assistance in a subsequent experiment than those not asked to imagine being part of a group.75 Additionally, “[w]hen sharp-shooters imagine shooting a gun, their entire body behaves similarly to actually doing so” (Barsalou, 1999, p. 588). Thus, empirical research indicates that merely imagining a scenario can subtly influence one’s perceptual experiences, decisions, and behavior.

Affective contagion occurs when imagining an emotionally charged situation causes us to feel or behave as though the situation were real, even when we are fully aware that it is fictional

(Gendler, 2006, p. 190). In a study by Harris et al. (1991), reported in Harris (2000), children were asked to imagine either a mean, horrible monster or a friendly rabbit occupying a visibly empty box placed in front of them. Despite knowing that they were merely pretending, several of the children who imagined a monster became frightened and asked that the experimenter not leave them alone with the box. Of all the children asked to imagine a creature who spent time alone with the box, half later claimed that they wondered whether there really was a creature inside it. Thus, merely imagining a creature led a substantial portion of the children to revise their assessment of the box’s contents, despite its being visibly empty when first examined

(Harris, 2000, p. 176). Contagion also predictably occurs in older subjects. Evidence indicates that exposure to fictional film and television violence leads to a short- and long-term increase in violent behavior in children, adolescents, and adults (Husemann & Taylor, 2006).

This is not to suggest that contagion is only (or even characteristically) a result of defective imagining. We tend to feel pity when reading Anna Karenina despite recognizing that the events of the story are completely fictional, and this is precisely what the story is designed to

75 Stenico & Greitemeyer (2014) report that subjects who played a video game with multiple characters were less likely to help others after the game concluded than subjects who played a game featuring just one other character.

94 do (Gendler, 2006). Usually, the point of imaginatively engaging with fiction is to produce a cognitive or affective response of some sort.76 As Gendler (2003, p. 134) notes, sexual fantasy provides a striking example of contagion: merely imagining a sexually arousing situation results in genuine sexual arousal, and this is by design.

Due to contagion, forming a mental image of a “slanty-eyed” person facilitates regarding the target as strange and exotic. By understanding the harm of slurs along these lines, we can make sense of their capacity to motivate derogatory thoughts and harmful discriminatory behavior in listeners. Likewise, contagion may account for the power of approbatives to influence us. Additionally, an embodied approach explains the power of intrapersonal uses of slurs and approbatives; uttering ‘slanty-eyed Chinese’ to oneself internally invites derogatory imagery just as uttering it out loud does.77 This approach also illuminates the varying intensity of different slurs (see Saka, 2007, p. 148, on variation among pejoratives). ‘Knuckle-dragging ape’ is more harmful than ‘tree hugger’ because of the different imagery that each invites; seemingly, forming a mental image of an African American person that depicts them as subhuman is potentially much more destructive than forming the more jocular image of someone showing affection toward a tree.

76 Walton (1978, 1990) argues that we feel at most “quasi-emotions” toward fictional characters and situations (e.g., quasi-fear of a fictional monster, which may involve physiological reactions, including an increased heart rate, muscle tension, etc., yet this is not genuine fear, since genuinely fearing something requires believing that it exists). Gendler and Kovakovich (2005) reply that we needn’t believe that an entity is real in order to have genuine, rational emotional responses to it. I will not take a stance on this issue. It would be enough for my purposes if imagining something could lead us to have genuine emotional reactions to real people and events. I take it that even if we can only quasi-pity Karenina, reading about her can dispose us to actually pity real people who lead similar lives. 77 One upshot of this is that the occurrence of the slur in the previous sentence and other pejoratives mentioned in this dissertation may have led readers to form an offensive image, despite the fact that these words are “mentioned” as opposed to “used.” This was not my intention. Nevertheless, I take it that even if some degree of sullying is inevitable when reading these words, this is a reasonable price to pay in getting clear on the ways in which they harm. Moreover, I am hoping that the pedagogical context in which these words are mentioned makes it easier to quarantine any unpleasant imagery that readers happen to entertain. 95

This is not to suggest that figurative, descriptive, and imitative slurs evoke imagery by definition – I have not attempted a conceptual analysis of the term ‘slur’. Rather, the suggestion is that to the extent that slurs are effective vehicles for spreading prejudice, imagery is explanatorily useful because of its contagion effects. As we have seen, embodied cognition research provides independent motivation for this approach. A potentially fruitful avenue of empirical research on slurs may involve testing for derogatory images specifically.

Imaginative resistance is one potential obstacle to forming derogatory imagery in response to figurative, descriptive, and imitative slurs. ‘Imaginative resistance’ refers to a phenomenon in which people resist engaging in a prompted imaginative activity (Gendler, 2000;

Liao et al., 2014). Hume (1985) held that asking people to adopt a moral point of view very different from their own provokes a kind of resistance:

a very violent effort is requisite to change our judgment of manners, and excite sentiments of approbation or blame, love or hatred, different from those to which the mind from long custom has been familiarized…. I cannot, nor is it proper that I should, enter into such [vicious] sentiments. (p. 247)78

Several authors have claimed that similar difficulties arise when we are invited to accept certain moral claims as fictionally true (see, e.g., Walton, 1994; Gendler, 2000). We find it difficult to imagine, e.g., that killing an infant is morally permissible simply because the infant is female, and one plausible explanation for this is that we think that infanticide for this reason would be impermissible in the actual world (Walton, 1994, p. 37). Liao et al. (2014) argue that the genre of a story often plays a role in triggering resistance. Liao et al. conducted a study in which subjects read a story in the style of a Greek myth. One character, Hippolytus, tricks another, Larissa, into binding herself to spend eternity with him. According to Liao et al. “audiences who are more

78 This quote from Hume was taken from Gendler (2000, p. 56). 96 familiar with stories in the tradition of Greek mythology found it easier to accept as fictional that it was morally right for Hippolytus to trick Larissa” (Liao et al., 2014, p. 350). Thus, imaginative resistance decreased as familiarity with the story’s genre increased. Liao et al. also had subjects read two nearly identical stories that differed only in genre. In the first, a contemporary police procedural drama set in the U.S., a woman, Mary, gives her infant child over to a preacher to be sacrificed because God demands it, whereas in the second story, set in an Ancient Aztec community, a woman, Ixchel, gives her child over to a priest to be sacrificed “in order to renew the sun” (Liao et al., 2014, p. 351). Subjects were asked whether they believed that Mary/Ixchel did the morally right thing in the story. Liao et al. report that, overall, participants did not accept that Mary did the right thing in the world of the fiction, whereas they tended to accept that Ixchel did the right thing in the ancient Aztech story.79 Liao et al. take these results to indicate that imaginative resistance varies depending on various contextual factors, including subjects’ familiarity with the genre, and features of the genre itself, which they take to be partially determined by factors such as the setting and time period of the story.

Resistance is not limited to propositional imagining. A large body of empirical literature indicates that the ability to form mental images and manipulate them varies depending on a variety of factors. Dror, Kosslyn, and Waag (1993) report that pilots performed better than non- pilots on a shape rotation task that required them to form a mental image of a certain geometric figure and rotate it. Ozel, Larue, and Molinaro (2004) did a similar study with athletes. They report that, regardless of their particular discipline, experienced athletes (those who train at least

10 hours per week) do better on mental rotation tasks than those who had never participated in a

79 Subjects were asked to give responses on a scale of 1 (strongly disagree) to 7 (strongly agree) to indicate whether they believed that the protagonist had done the morally right thing in the story. Liao et al. report a mean score of 3.03 with a standard deviation of 2.173 for the police procedural drama involving Mary, and a mean of 4.17 with a standard deviation of 2.422 for the Aztec story involving Ixchel (p. 352). 97 sporting competition and those who participate in sports as a leisure activity (2 hours or less per week). Other studies indicate that the ability to imaginatively simulate certain experiential states varies among people of different backgrounds. Olsson et al. (2008) asked both experienced and novice high jumpers to imagine performing a jump from an “internal” perspective (i.e., to imagine what it feels like to perform the jump, and not to imagine watching someone else do it).

An fMRI scan revealed that imaginatively simulating the experience of a high jump led to the activation of motor regions of the brain in experienced jumpers. However, novices who attempted the same imaginative exercise activated only visual and parietal regions of the brain

(Olsson et al., 2008).

In light of the empirical data on imaginative resistance, it is plausible that slurs trigger different levels of resistance in different hearers or readers. We should expect hearers who do not harbor negative attitudes about a target group to have difficulty forming derogatory imagery when hearing slurs for the group in question. Perhaps previous exposure to humanizing depictions of targets leads to resistance in people who overhear slurs. Similarly, it is plausible that real life interaction that requires cooperating with targets and seeing them as social equals may make forming unflattering imagery more difficult.80 If one does not form an image, contagion cannot occur. Conversely, an increase in prejudice should lead to a decrease is resistance. Features external to the agent responding to slurs may also be relevant. Taking one’s nation to be participating in a just war against the target may make forming an image that depicts them as cruel, remorseless killing machines easier (the so-called War on Terror may be one such conflict). We may also discover analogues to the genre differences that lead readers to accept

80 In a meta-analysis of 515 studies, Pettigrew and Tropp (2006) found “substantial evidence that intergroup contact can contribute meaningfully to reductions in prejudice across a broad range of groups and contexts” (p. 766). See also Turner and Crisp (2010), who report that merely imagining positive intergroup contact leads to a reduction in both explicit and implicit prejudice. 98 certain moral propositions as fictionally true in the context of some stories, but not in others.

Using a slur in the course of telling a joke may provoke less resistance than using the same slur in other contexts. Of course, these are all empirical questions (I know of no studies that conclusively confirm or disconfirm these predictions). However, given that, in general, a subject’s ability to engage in propositional imagining, to form visual imagery, and to simulate experiential states varies depending on various contextual factors (including the subject’s background knowledge, experiences, and features of the imaginative exercise the subject is attempting to undertake), we should expect that a hearer’s ability to imaginatively respond to slurs will vary based on contextual factors.

It is possible to be affected imagistically by slurs even when we are not consciously aware of forming imagery. Schwitzgebel (2002) notes that people’s self-reports about their experience of visual imagery often vary widely. Some report highly vivid images that are nearly identical to perceptual experiences of objects in the external world, while others report difficulty forming imagery, yet these self-reports do not correlate with performance on tests that psychologists suppose require mental imagery, such as mental rotation tasks, tests of visual creativity, and tests of visual memory (Schwitzgebel, 2002, p. 43). Further, a large body of empirical research on weapons bias (the phenomenon whereby subjects tend to misidentify harmless objects as weapons) indicates that images can affect us without our awareness. If this is correct, imagery may account for the experimental results of Carnaghi and Maass (2007), who report that exposure to anti-gay slurs activates implicit prejudice in subjects who identify as heterosexual, even if we suppose that the subjects were not consciously aware of forming relevant imagery. Payne (2006) reports that priming subjects with an image of an African

American’s face triggered weapons bias. Seemingly, perceiving the image activated implicit

99 prejudice even when it was not processed consciously. Additionally, research on blindsight indicates that the neural processing underlying perception and the formation of images can occur with or without conscious experiences (Barsalou, 1999, p. 583). If hearing or reading slurs activates certain neural processes that underlie objectionable imagery or other relevant experiential states, then it may be possible for these words to influence us without generating conscious experiences (and so slurs may sully us without making us aware of experiences that lead us to view targeted groups in a derogatory light). However, this is consistent with competent hearers sometimes getting conscious images when exposed to slurs that make them feel sullied.

5. Advantages of a Pluralistic Approach over Standard Theories

Most theorists hold that slurs are derogatory in virtue of the contents they convey.

Chapter 2 established that standard semantic and pragmatic theories have yet to explain how semantically expressing, implying, or presupposing a proposition about a target influences listeners. Hom and May offer no account of how ascribing the property of being worthy of negative evaluation to a target motivates listeners to think along these lines. Likewise, implicature theorists have yet to develop a notion of implicature that explains how implying a derogatory statement could motivate others to think similarly. The presupposition theory is also deficient; there is a rather large gap between allowing a proposition about the target group into common ground, and actually having disparaging thoughts about them (Richard, 2008, p. 21). In general, attempts to give an analysis of slurs using models of rational, cooperative communication do not capture the phenomenon at issue. We also saw in Chapter 2 that expressivists face analogous worries. Given that slurs constitute a heterogeneous group, we should not treat them as constituting a semantic or pragmatic kind.81 The pluralist view

81 Nunberg (forthcoming) makes a similar claim about slurs. 100 developed above avoids the problems that beset other theories. For many slurs, imagery provides the missing ingredient that other theories have failed to supply. As empirical data indicates, mental images have the power to influence us in ways that are not captured by other mechanisms

(like implicature, presupposition, etc.). This is not to deny that one may imply, presuppose, or express a variety of attitudes with slurs; rather, it is simply to claim that imagery does important explanatory work that these mechanisms do not.

While Camp provides a novel approach (perspectivalism), her view fails to provide a satisfactory account of slurs’ rhetorical power. Recall that for Camp (2013), when a speaker utters a slur, she signals her derogatory perspective by virtue of the slur’s conventional meaning.

For Camp, the fact that slurs’ signal derogatory perspectives accounts for their rhetorical power.

The broadly Davidsonian view I have defended here does not hold that slurs signal anything about the speaker’s perspective in virtue of their conventional meaning. Further, Camp (2013) appeared to be making a very strong claim: that merely interpreting a slur requires buying into

(or being disposed to buy into) the perspective it signals, and this is why slurs generate complicity in hearers. In a forthcoming paper on metaphorical insults, Camp claims that merely comprehending metaphors requires employing a kind of perspectival thinking. She holds that in using a metaphor during a conversation, the speaker requires her interlocutor to use a characterization (that is, a cluster of features) of one thing to structure a characterization of another thing in order to comprehend her utterance. In interpreting ‘Juliet is the sun’, e.g., the hearer takes a cluster of the most prominent features of the sun and identifies relevant matches to features within the subject characterization, i.e., Juliet (Camp, forthcoming, p. 5). Camp claims that “framing produces complicity because merely understanding the utterance requires a hearer to mold his mind in the speaker’s image: to structure his overall thinking so that the relevant

101 features really are intuitively prominent and central for him” (Camp, forthcoming, p. 5). I am not attributing any such feature to slurs. As noted above, slurs may fail to influence us because of imaginative resistance. A competent hearer may know precisely what sort of imaginative exercise a slur invites, i.e., she may comprehend the speaker’s utterance, and thereby understand the sort of derogation it is standardly used to accomplish, without complying. Further, even if one forms an image quickly and spontaneously in response to overhearing a slur or metaphorical insult, one need not endorse the image one forms, and hearers who form imagery may quarantine it from their other attitudes. A non-racist who mentally forms a caricatured image of an East

Asian person upon hearing ‘slanty-eyed’ will likely see the image as a perverse and vicious fiction, and as a result, quarantine it from her other representations. In that case, contagion will not occur.82 Thus, pace Camp, culturally and linguistically competent listeners can comprehend a slurring utterance without getting, or being disposed to get, a derogatory perspective, and moreover, those who form imagery and quarantine it need not be complicit in any kind of derogation (or any other wrongdoing for that matter). However, as we have seen, imagery may not be subject to conscious awareness, and so hearers may be unable to consciously step back from an image and disavow it. Given what empirical research on resistance and quarantining indicates, hearers who are not explicit bigots, yet are not subject to imaginative resistance and have difficulty quarantining imagery, are likely to become prejudiced through contagion as a result of being exposed to imagistic slurs.

Another worry is that Camp’s view has counterintuitive implications regarding our moral obligations. Imagine a non-native English speaker, Jane, who is not competent with slurs in

82 Quarantining and imaginative resistance may explain why anti-gay slurs did not activate prejudice in subjects who identify as homosexual in Carnaghi and Maass’s (2007) study. One possibility is that thinking of oneself as a member of the target group makes it easier to quarantine anti-gay imagery. Another is that anti-gay slurs trigger resistance in listeners who identify as homosexual. 102

English. When Jane overhears utterances of a slur in English, she is not complicit in derogation because she lacks comprehension. If, as Camp holds, merely comprehending the utterance of slurs is sufficient for making hearers complicit in derogation, then presumably Jane has a moral obligation to never become competent with slurs. Jane ought not look up any slurring words in a dictionary, nor should she take part in conversations that may lead her to have minimal linguistic competency with slurs. Suppose Jane is an academic. In that case, she ought not referee papers on slurs or supervise a dissertation on them. For if a speaker who is not already competent with slurs were to do any of these things, she would thereby open herself up to being complicit in derogation. Since it is highly implausible that we have a moral obligation not to do any of these things, Camp is wrong about slurs’ power to make us complicit.83

5. Appropriation

The embodied approach developed here has the resources to explain the phenomenon of appropriation. While I will not provide a full account here, I will offer some speculative remarks about how some slurs may be appropriated. As noted in Chapter 2, in its most general use

‘appropriation’ refers to the repurposing of something. As numerous theorists observe, slurs are often appropriated by targets in order to eliminate or lessen their derogatory power (Brontsema,

2004). This can happen in at least three ways. First, new imagery may replace old, unflattering imagery. Hom (2008, p. 428) gives the following example of appropriation by rapper Tupac

Shakur from the documentary film Tupac: Resurrection (2003): “Niggers was the ones on the rope, hanging off the thing; Niggaz is the ones with gold ropes, hanging out at clubs.” Shakur’s comments could be taken to suggest that in some appropriated uses ‘Nigger’ invites an image that depicts African Americans as victims of racism, yet when stylized as ‘Nigga’, the term

83 I am indebted to Daniel Silvermint for helpful discussion on this point. 103 invites an image that depicts African Americans as people who are wealthy, successful, and empowered.84

In a second sort of appropriation, a speaker uses a slur in an ironic way. Paralinguistic features of the utterance of a slur, such as a mocking or distancing tone, may indicate to hearers that the speaker is not inviting an image, but is instead attempting to ridicule people who do.85

By making-as-if to invite an image in this way, one can mark the image in question as a perverse, vicious fiction.

Finally, appropriation may be facilitated by the loss of a slur’s imagery over time.

Perhaps at one point in the twentieth century, ‘queer’ invited an image that depicted its target as a sexual deviant. In any event, almost no one finds contemporary uses of the term offensive.

Now ‘queer’ admits of a variety of non-derogatory uses, e.g., in academic contexts (‘queer theory’, ‘queer studies’). Once the slur’s previous imagery faded (assuming there ever was any), and as a result its use was no longer objectionable, the term became a neutral way of referring to

LGBT people.

6. Conclusion

The current consensus is that slurs are derogatory in virtue of the contents they convey – what they imply, signal or otherwise express. If this is correct, an adequate account of the rhetorical effects of slurs – in particular, their capacity to motivate derogatory attitudes in

84 Hom (2008) suggests an interpretation like this, though he holds that in appropriation, a change occurs at the level of the slur’s truth-conditional content. The story I’ve give above is consistent with Hom’s – it may be that in some cases of appropriation, a slur’s new semantic content introduces new constraints on how hearers are to imaginatively respond when speakers use the term. 85 Croom (2013) notes that in uttering ‘nigger’ as ‘nigga’, a speaker implements a phonological rule of African American Vernacular English, namely deleting /r/ when it is preceded by a vowel. He claims that in doing so, out-group speakers distance themselves from derogatory attitudes standardly expressed by ‘nigger’, and that as long as speakers are sincere, their use of the term is non-derogatory. However, it is not clear that uses of this sort by white speakers are entirely innocuous even when such speakers are sincere. 104 listeners – requires an explanation of how the posited derogatory content causes a cognitive or affective change in hearers. The approach developed here is non-cognitivist in the way that

Davidsonian accounts of metaphor are non-cognitivist, namely it explains the sullying effects of slurs by appeal to imagery they invite (though I have not endorsed everything Davidson has said about metaphor). Imagery has explanatory purchase here because of its contagion effects. How an image ultimately affects us depends crucially on what we do with it – it matters whether we savor it and add details, or filter it out immediately, though these processes may not always be conscious or deliberate. This is not to claim that all slurs harm by inviting imagery. Since slurs are a heterogeneous group, a pluralistic theory is needed. Slurs may harm listeners in a visceral way that does not crucially rely on imagery. Of course, slurs may perpetrate both kinds of harm – they may both sully and shock (or offend, insult, etc.). It is always possible for a slur to transition over historical time from being a vehicle for inviting vivid imagery to being merely a prohibited term for the target group to a neutral, non-offensive classifying term (and vice-versa).

105

Chapter 4: Do Racists Speak Truly? On the Truth-Conditional Content of Slurs

1. Introduction

Are statements like ‘Native Americans are redskins’ and ‘East Asians are slanty-eyed’ literally true? Most theorists hold that each slur has a neutral counterpart, a term that denotes the slur’s target group without derogating them, e.g., ‘Jewish’ is a neutral counterpart to ‘kike’

(Hornsby, 2001, p. 129). According to a widely accepted view, which I shall call Neutral

Counterpart Theory (hereafter NC), the literal, truth-conditional content of a slur is identical to the literal, truth-conditional content of its neutral counterpart. Gibbard (2003) aligns himself with

NC when he claims that non-prejudiced speakers can recognize predications of slurs to targeted individuals as “truths objectionably couched” (p. 300). Although NC is prima facie plausible as a thesis about single-word noun-phrases, such as ‘chink’ and ‘dago’, it does not extend to

‘redskin’, ‘slanty-eyed’, ‘curry muncher’, ‘Jewish American Princess’, and ‘camel jockey’, nor does it extend to ‘ching chong’, which does not encode any truth-conditional content whatsoever, and instead denigrates its target by virtue of its iconicity – or so I shall argue in the present chapter.

As we have seen, acts of expression are distinct from vehicles of expression (Bar-On,

2015). When used as a transitive verb ‘slur’ refers to the act of denigrating individuals qua members of a certain group, and when used as a noun ‘slur’ refers to symbolic vehicles that speakers use in performing such acts (Anderson & Lepore, 2013b, p. 351; Hom & May, 2013, p.

106

294). One needn’t use a conventional slurring word to slur a target. As Jeshion (2013b, p. 315) observes, a speaker could slur someone by uttering a neutral term for their group in a denigrating tone, as in

(1) Yao is ChineseC where the superscript ‘C’ indicates a contemptuous tone. My aim is to undermine the view that all conventional slurring words and their neutral counterparts are truth-conditionally equivalent.

2. Proponents of NC

Numerous authors commit themselves to NC, yet not all agree on what gives slurring words their power to derogate. Frege was perhaps the first NC theorist. According to Frege,

‘horse’, ‘steed’, and ‘nag’ all have the same sense and reference (1956, p. 295). On an extension of this view, each slur has the same sense and reference as its neutral counterpart. Some NC theorists appeal to Grice’s notion of conventional implicature. Grice (1989) distinguished the literal meaning of an utterance from what a speaker means by it, and observed that speakers often mean more than what they literally say. As proponents of the CI view of slurs (discussed in

Chapter 2), Potts (2005), Williamson (2009), McCready (2010), Whiting (2013), and Lycan

(2015) hold that slurs encode the same truth-conditional content as their neutral counterparts, yet the use of a slur conventionally implies a certain negative attitude about the target group.

Some expressivists are committed to NC. Jeshion (2013b) holds that slurring words are expressive of contempt. On her view, the truth-conditional component of a slur serves only to fix its extension, and is separate from the expressive component (Jeshion, 2013b, p. 316).86 As we saw in Chapter 2, for Jeshion, derogatory uses of a slur for group G express contempt for Gs on

86 While Saka (2007) defends a hybrid expressivist theories of slurs, he argues that we are ignorant of the truth-conditions of slurring statements (see Chapter 2 of his 2007). 107 account of their group membership, but not by virtue of any semantically encoded descriptive content, such as despicable because of being members of G (Jeshion, 2013b, p. 317).

Finally, Anderson and Lepore (2013a, 2013b) defend a deflationary theory of slurs

(discussed in Chapter 2) which is committed to NC. The only difference between slurs and their neutral counterparts, on their view, is that the former are prohibited. In opposition to implicature theorists and expressivists, Anderson and Lepore claim that “slurs are prohibited words not on account of any content they get across, but rather because of relevant edicts surrounding their prohibition” (2013a, p. 26).

3. Recent Challenges to NC

Recently, several theorists have challenged NC, yet their attempts to undermine the view have not been successful. As noted in Chapter 2, Hom (2008) holds slurs are not intersubstitutable salva veritate with their neutral counterparts in intensional contexts. Suppose A says, ‘Yao is a chink’, and B attempts to report on A’s utterance with ‘A said that Yao is

Chinese’. Since B’s indirect report is incomplete, the derogatory content of A’s utterance must be part of what A said, and so ‘chink’ and ‘Chinese’ are not truth-conditionally equivalent (Hom,

2008, p. 425).

NC theorists may respond that B’s report was accurate, yet she failed to note the derogatory way in which A categorized Yao. One could imagine B specifying the derogatory component of A’s utterance in a way that is consistent with NC: ‘A said that Yao is Chinese, though, being a foul-mouthed racist, A didn’t put it quite that way’ (Whiting, 2013, p. 374).

Thus, Hom’s data on the failure of intersubstitutability of slurs and their neutral counterparts in

108 indirect reports does not provide compelling evidence against NC.87,88

Another objection due to Hom is that any theory according to which slurs are synonymous with their neutral counterparts generates the unintuitive result that racist claims like

‘Chinese are chinks’ are analytically, and therefore, necessarily true (2008, p. 424). Given that

‘Chinese’ and ‘chink’ are truth-conditionally identical, ‘Chinese are chinks’ has the same truth- conditions as ‘Chinese are Chinese’. However, Whiting (2013) responds that when one claims that ‘Chinese are chinks’ is true, one might be endorsing only what is said (namely, that Chinese are Chinese), and not how it is said. As Whiting observers, we might feel squeamish about claiming that such sentences are true because claiming that a sentence is true is often a way of endorsing its use. For NC theorists, however, there is nothing racist about what is literally claimed by a speaker who utters ‘Chinese are chinks’; rather, what is racist is what the speaker is doing – conventionally implying or otherwise expressing – in uttering this sentence.89 Thus,

Hom’s objection trades on an -ing/-ed ambiguity (Whiting, 2013, p. 373). Additionally, Sennet and Copp (2015) appeal to the notion of ‘semantic charity’ in order to show that ‘Jews are kikes’

87 When I submitted a version of this chapter to the journal Thought, an anonymous referee pointed out that another way in which NC theorists could resist Hom’s intersubstitutability objection is by arguing that slurs or their neutral counterparts are what Richard (1993) calls “articulated terms.” According to Richard, even if we assume that two terms a and b are directly referential and co-referential (and so they contribute the same semantic values to sentences in which they occur), a might contribute an articulation of the referent to what is said that b does not, and so the two terms may not be intersubstitutable salva veritate. For Richard, ‘[the proposition] that mathematics reduces to logic’ is articulated, while ‘logicism’ is not, and thus, these terms are not intersubstitutable. Analogously, NC theorists could hold that a slur and its neutral counterpart may not be intersubstitutable when one is articulated and the other is not. However, the proposed solution would not account for failure of intersubstitutability for ‘spic’ and its counterpart, ‘Hispanic’, and ‘dago’ and its counterpart, ‘Italian’, since none of these terms are articulated. 88 Hom and May (2013, p. 304) provide additional data designed to undermine NC that involves substitutions of slurs with neutral counterparts in contrasting sentence pairs, e.g., “I’m not a kike, but I am a Jew (True)”; “I’m not a Jew, but I am a Jew (False).” In response, Whiting (2013, pp. 370-72) suggests a metalinguistic reading of sentences like the former, on which the speaker is claiming that she should not be called ‘kike’. 89 Deflationists differ from other NC theorists on this point. Utterances of ‘Chinese are chinks’ are offensive, for deflationists, not because they express derogatory content, but because they reference Chinese people in a prohibited way. 109 is analytically true, contra Hom. They argue that since ‘kike’ is commonly used by anti-Semites to refer to Jews, and since non-anti-Semites would generally recognize that anti-Semites who use the term are attempting to refer to Jews, the general pattern of use makes Jews the best fit for the extension of ‘kike’.

Richard (2008) raises an objection to NC that is similar to Hom’s:

Imagine standing next to someone who uses S as a slur. …[T]he racist mutters That building is full of Ss. Many of us are going to resist allowing that what the racist said was true. After all, if we admit its truth, we must believe that it is true that the building is full of Ss. And if we think that, we think that the building is full of Ss. We think, that is, what and as the racist thinks. (2008, p. 13)

Richard seems to be guilty of the same kind of confusion we saw in Hom. Suppose that ‘S’ is a slur for Italians. If NC is correct, what the racist said is nothing more than that building is full of

Italians. Certainly, there is nothing derogatory per se about thinking that a building is full of

Italians. For implicature theorists and expressivists, one could coherently avow the speaker’s claim about the ethnicity of the building’s occupants and simultaneously disavow the derogatory attitudes expressed in making that claim (or, if one is a deflationist, disavow the use of a prohibited term for the target) – either internally or out loud. Thus, pace Richard, admitting the truth of what the racist said does not entail thinking as the racist thinks.90

A fourth objection is due to Hom and May (2013). They give the following “modal- conceivability argument,” whose conclusion is inconsistent with NC (Hom & May, 2013, p.

304):

(2) It is conceivable for there to be Jews without Kikes.

(3) Whatever is conceivable is possible.

90 Hom and May (2014) offer additional objections to NC that are distinct from those I give below. Sennet and Copp (2015) provide rejoinders to Hom and May. 110

(4) Therefore, it is possible for there to be Jews without Kikes.

Of course, (4) entails that the meaning of ‘Jew’ and ‘Kike’ differ. So, if this argument succeeds,

NC is false. I will grant (3) and focus on the controversial premise in (2). The motivation for (2), according to Hom and May, is that we can imagine a world like ours that is “morally perfect,” and thus “devoid of racism of any kind” (2013, p. 304). In such a world, they argue, “it is clearly conceivable for there to be Jews without Kikes” (Hom & May, 2013, p. 304). Notice that the argument goes through only if we read the occurrences of the slur in (2) and (4) as metalinguistic, so that (2) says that it is conceivable for there to be Jews who are not called

‘kikes’, and the conclusion in (4) says that it is possible for there to be Jews who are not called

‘kikes’.91 However, this reading is consistent with NC, and does not support Hom and May’s view that the truth-conditional content of slurs differs from that of their neutral counterparts.

Also, Sennet and Copp (2015, p. 1090) suggest that Hom and May have confused what a term means in the actual world and when applied across different worlds with what it means for people in those other (merely possible) worlds. Perhaps ‘kike’ literally means something other than ‘Jew’ in the mouths of speakers in worlds free of anti-Semitism. This does not establish the truth of (2). However, as I explain below, a revised version of Hom and May’s conceivability argument that uses different slurs is more plausible.

A fifth objection is due to Glezakos (2012), who challenges the view (adopted by

Hornsby and others) that we can draw a distinction between slurs and purportedly “neutral” counterpart terms, such as ‘Black’ and ‘Asian’ (see also Ashwell, 2016). She argues that alleged neutral counterpart terms are in fact either negatively or positively valenced due to social facts about how the groups referenced by these words are treated. She points to terms such as

91 As Sennet and Copp (2015, p. 1091) note, a non-anti-Semitic world is clearly conceivable, and presumably, no one in such a world would use ‘kike’ if it was a vehicle for denigrating Jews in that world. 111

‘Chinaman’ and ‘Oriental’, which English speakers in the early twentieth century took to be neutral classifying racial terms, yet they are now widely held to be derogatory, and suggests that we are making a similar error in thinking of terms such as ‘Black’, ‘White’, ‘Asian’, and

‘Woman’ (among others) as neutral. Glezakos claims that

when it is indeed a social fact that members of a particular group are, in virtue of their membership, subject to discrimination, restriction, or violence, the meaning of words by means of which the group and its members are specified will reflect the particular disfavor in which they are held. …[W]hen the practices of classification from which a word obtains its meaning are marked by non-neutral assessment and attribution, we cannot maintain that the word is fundamentally different from the ones that we identify as derogatory, even if we find it more suitable in certain ways. (p. 390)

Glezakos suggests that the use of racial classifying terms in statements such as (5)-(7) indicate that these terms are evaluative:

(5) Applications from Black students are especially encouraged.

(6) Asians are good at math.

(7) Whites only.

Glezakos does not offer an account of the literal, truth-conditional meaning of group- classifying terms like ‘Black’, ‘Asian’, and ‘White’; her aim is merely to show that terms such as these are not neutral, and so it is a mistake to draw a distinction between slurs and neutral counterparts. If this is correct, and NC relies on the existence of neutral counterparts to determine the truth-conditional meaning of slurs, then NC is in trouble. However, NC theorists may respond that even if Glezakos is correct that the racial classifying terms she mention are not neutral, this needn’t pose a problem for NC, since they need only commit themselves to the possibility of neutral classifying terms. Glezakos may hold that group classification is itself an

112 intrinsically evaluative act, and in that case neutral racial classifying terms would be ruled out in principle, and so any term one could invent for a group would be necessarily evaluative. Yet this would be an extreme and implausible view. Suppose that ‘Italian-American’ is a non-neutral term for a person’s ethnicity or nationality. We can imagine a term t that picks out the same group denoted by ‘Italian-American’, yet t’s conventional function is not to evaluate the group in question in any way whatsoever. Suppose that t was introduced by alien social scientists who simply wanted to classify a specific group of Earthlings, and that these aliens have no connection whatsoever to a history of discriminatory or preferential treatment of the group under consideration. Even if racial classification is often an evaluative enterprise in that it tends to be motivated by certain pernicious evaluative beliefs and practices, we have no reason to think that it must be. As far as the present example is concerned, NC theorists are committed to holding only that the literal, truth-conditional content of slurs that target Italian-Americans is identical to the literal, truth-conditional content of t. Whether a term like t actually exists does not bear on the truth of NC in any way whatsoever.

One major problem with Glezakos’s argument is that she equivocates on the term

‘evaluative’. A word may be evaluative because a speaker uses it to evaluate someone, yet this is consistent with the word’s conventional function being merely to classify. Theorists could hold that while ‘Black’, ‘Asian’, and ‘White’ are not evaluative terms, a speaker who uses these term when uttering (5)-(7) evaluates the relevant groups in a positive or negative way, yet this is not due to the meaning of the racial classifying terms she is using. Seemingly, the contribution by the words ‘encouraged’, ‘good’, and ‘only’ in (5)-(7) make it the case that these are evaluative statements. Thus, it is puzzling that Glezakos takes these statements to provide support for her claim that racial classifying terms are themselves evaluative. In any event, NC does not predict

113 that terms for a person’s race cannot be used in acts of evaluation, and so the data that Glezakos provides should not lead us to reject NC or, for that matter, any view that distinguishes slurs from neutral counterparts.

4. Some New Data

Although proponents of NC can adequately respond to the challenges raised by Hom and

May, Richard, and Glezakos, NC does not extend to all slurs.92 Let us assume that ‘East Asian’ is a neutral counterpart to ‘slanty-eyed’. Let us also suppose that ‘East Asian’ as a neutral term for various groups including, inter alia, Chinese, Japanese, Koreans, Mongolians, and Vietnamese, all of whom constitute the target group for this slur.93 If ‘slanty-eyed’ is a slur, NC theorists are committed to holding that (8a) and (8b) are truth-conditionally identical:

(8a) Mark is slanty-eyed.

(8b) Mark is East Asian.

If, as I shall argue, (8a) and (8b) differ in their literal truth-conditions, this is due to a difference in the truth-conditional content that ‘slanty-eyed’ and ‘East Asian’ encode. NC also fails to extend to ‘curry muncher’, a slur for various South Asian groups. It is implausible that the literal, truth-conditional content of ‘curry muncher’ is identical to that of its neutral counterpart, ‘South

Asian’ (or whatever we take the appropriate neutral counterpart to be), since one could be a

92 Jeshion (2013b) intends for her expressivism to cover only slurs that are truth-conditionally equivalent to their neutral counterparts, such as ‘kike’. She claims that ‘Christ killer’ and ‘jungle bunny’ carry descriptive content not encoded by their neutral counterparts (though she does not argue for this claim), and so these slurs, as well as metaphors like ‘Jewish American Princess’ and adjectival slurs like ‘Chinky-eyed’, are “largely outside the purview of [her] investigation” (2013b, p. 311). Given that Jeshion intends for her view to cover only a limited set of slurs, the objections I give here will not apply to it. However, as I argue below, Jeshion’s move would not make for a good general defense of NC, since proponents may not simply cordon off examples that do not fit their theory. 93 Nothing crucially relies on the assumption that ‘East Asian’ is a neutral term for these groups. If we have reason to believe that this term is not neutral, we could simply invent a term designed to classify the relevant groups in a way that is not positively or negatively valenced. 114 member of the slur’s target group without ever having eaten curries.

One major problem with NC is that it predicts that the following racist utterances will be absurd when read literally, yet they seem intelligible:

(9) Mark was once slanty-eyed, but he got plastic surgery to change that.

(10) Mira used to be a curry muncher, but hasn’t been one since she changed her diet.

Suppose that Mark and Mira are of East Asian and South Asian ancestry, respectively. Read literally, (9) is an intelligible (albeit racist) claim, yet for NC theorists (9) is absurd. If ‘slanty- eyed’ is a slur, NC theorists are committed to holding that being literally slanty-eyed is just a matter of being of East Asian ancestry, yet this is not a feature one could change by means of surgery.94 Likewise, if ‘curry muncher’ is a slur and NC is correct, on a literal reading (10) says that Mira used to be of South Asian ancestry, but this is no longer the case due to her changing her diet. Of course, changing one’s diet could not literally alter one’s race. Thus, NC theorists must hold that (10), which seems intelligible (though racist), is absurd when read literally. The problem seems to be NC theorists’ assumption that the literal meaning of ‘slanty-eyed’ and

‘curry muncher’ is identical to the literal meaning of their respective neutral counterparts.

NC theorists may respond that being literally slanty-eyed is a matter of meeting two conditions: (i) being of East Asian ancestry, and (ii) exhibiting a certain perceptible feature, such as having epicanthic folds (i.e., folds of skin covering the inner angle of each eye). In that case,

(9) is no longer absurd, since it could be literally true that Mark is of East Asian ancestry and had a fold of skin over the inner angle of each eye which was surgically removed. Similarly, NC theorists might hold that being a literal curry muncher is a matter of (i) being of South Asian

94 A similar argument could be given using ‘slant-eyed’ or ‘slant-eye’, which functions grammatically as a noun. 115 ancestry, and (ii) being someone who tends to consume curry vigorously. In that case, (10) is no longer absurd when read literally. The trouble is that ‘slanty-eyed’ and ‘curry muncher’ target all people of East Asian and South Asian ancestry, respectively, yet having a fold of skin over the inner angle of one’s eyes is not a necessary condition on being of East Asian ancestry,95 and eating curry vigorously is not a necessary condition on being of South Asian ancestry. Thus, under the current proposal, ‘slanty-eyed’ and ‘curry muncher’ could have suitable neutral counterparts whose truth-conditional content differs compared with what their corresponding slurs encode. So, augmenting the truth-conditional content of slurs in the suggested way would not provide succor for NC.

Recall the modal-conceivability argument by Hom and May (2013). While this argument did not succeed in showing that ‘Jew’ and ‘Kike’ differ in their truth-conditional content, it looks more plausible when we plug in descriptive slurs, as in the following statements:

(11) It is conceivable for there to be East Asians without there being slanty- eyed people.

(12) It is conceivable for there to be Native Americans without there being redskins.

If we accept (11) and (12), and we grant the assumption that whatever is conceivable is possible, then it follows that it is possible for there to be East Asians without slanty-eyed people, and possible for there to be Native Americans without redskins. The motivation for (11) and (12) is not that we can imagine a world in which there is no discrimination toward East Asians or Native

Americans.96 Rather, the idea is that it is possible for East Asians and Native Americans to lack

95 Zack (2002, pp. 87-8) argues persuasively that there is no scientifically respectable notion of race according to which being a member of a certain race depends crucially on any phenotype. 96 Notice that we need not read the occurrences of the slurs in (11) and (12) as metalinguistic (so that they say that we can conceive of a world where no one is called ‘slanty-eyed’ or ‘redskin’). 116 whatever perceptible feature racists attribute to them. We needn’t settle the question of what those features are here. Given that phenotype does not determine race, for any particular phenotype, we can imagine East Asians and Native Americans lacking it.

In addition to descriptive slurs, figurative slurs appear to be problematic for NC. Let us assume that ‘Jewish American female’ is a neutral counterpart to ‘Jewish American Princess’.

(13a) and (13b) appear to have different literal truth-conditions:

(13a) Alex is a Jewish American Princess.

(13b) Alex is a Jewish American female.

If (13a) is literally true, Alex is a Jewish American daughter of a monarch. Of course, on most occasions of use the speaker would not intend (13a) to be taken literally; rather, the speaker intends to convey that Alex is spoiled, self-centered, and materialistic because of her ethnic background (Jeshion, 2013a, p. 318). Since being literally female is not the same as being a literal Princess, ‘Jewish American Princess’ does not have the same literal, truth-conditional content as its neutral counterpart.

In general, live metaphorical slurs pose a problem for NC, since their literal meaning is necessarily distinct from the literal meaning of their neutral counterparts. Despite widespread disagreement over how metaphors work, nearly all theorists agree that “getting” a metaphor is a matter of interpreting an utterance’s linguistically encoded meaning non-literally (Camp &

Reimer, 2006). Imagine a racist speaker derogating African Americans by calling them ‘knuckle- dragging gorillas’. Before the metaphor can bring about its characteristic rhetorical effect on hearers, they must first process its literal meaning. If the slur’s literal meaning were identical to the literal meaning of its neutral counterpart (‘African American’), we would have no way of

117 arriving at the racist metaphorical content of the speaker’s utterance.97

One possible line of response is that above counterexamples are actually idioms that no longer retain their original compositional meaning.98 Many theorists (e.g., Chomsky, 1980) assume that idioms are non-compositional in that their meanings are not a function of the meanings of their parts and how they are combined. ‘Kick the bucket’, e.g., is semantically unstructured such that its meaning is not distributed over its parts (Nunberg, Sag, & Wasow,

1994). NC theorists may hold that the above counterexamples are similarly non-compositional.

However, as (9) and (10) illustrate, ‘slanty-eyed’ and ‘curry muncher’ are derogatory when used literally. This suggests that these terms are not figurative or idiomatic slurs. Regardless, not all idioms are non-compositional. Nunberg, Sag, & Wasow distinguish idiomatic combining expressions, whose parts have meanings that combine to form the meaning of the entire phrase

(e.g., in ‘pulling some strings’, ‘strings’ refers to personal connections, while ‘pulling’ refers to the act of exploiting those connections) from idiomatic phrases like ‘kick the bucket’. One motivation for this distinction is that some idioms can be modified and extended (e.g., ‘leave no legal stone unturned’, ‘Brazil is a team with more rabbits in its hat than most’), yet if all idioms are non-compositional, this would not be possible (Nunberg, Sag, & Wasow, 1994; Egan, 2008).

As (14)-(16) illustrate, some slurs are modifiable and extendable in racist statements:

(14) He’s half-Chinese, so his eyes are somewhat slanted.

(15) Those camel jockeys have traveled a long way, so their rides must be thirsty.

(16) They’ve stayed out of the sun for a few generations, so they’re not quite as redskinned as they used to be.

97 Non-cognitivists may balk at the phrase “metaphorical content.” For them, metaphors have no content other than their literal content (see, e.g., Davidson, 1978; for criticism, see Camp & Reimer, 2006). Since non-cognitivism is not amenable to NC anyway, I will not discuss it in detail here. 98 Thanks to William Lycan for this suggestion. 118

Thus, even if the slurs in (14)-(16) are idioms, it does not follow that their meanings are determined non-compositionally.

Finally, NC does not extend to slurring words that denigrate by virtue of their iconicity.

‘Ching chong’ targets Chinese people, yet it does not seem to encode any truth-conditional content whatsoever.99 Notice the difference between yelling ‘Ching chong!’ while passing a

Chinese person, and yelling ‘You’re a chink!’, which seems to encode truth-conditional content.

This is not to suggest that all slurs lack neutral counterparts, as Glezakos aims to show. I will assume that most slurs have (or could have) neutral counterparts, since I aim only to show that not all slurs have the same truth-conditions as their neutral counterparts.100

It might be argued that the objections raised so far rely on an equivocation. NC theorists could appeal to a distinction between acts of slurring and slurring words. Although the purported counterexamples may constitute acts of slurring, they are not conventional slurring words, so the reply might go. Consider the following exchange:

(17) A: John can’t hold a job.

B: Well, he is white.

While we may accuse B of slurring John (and other white people) by means of an implicature, B did not use any slurring words.101 NC theorists could argue that while speakers may use ‘slanty- eyed’, ‘curry muncher’, ‘knuckle-dragging gorilla’, etc. to slur targets, these are neutral phrases that do not derogate by convention. Additionally, theorists could hold that ‘ching chong’ is a

99 Similarly, ancient Greeks used ‘bar-bar’ (the ancestor of ‘barbarian’) to imitate foreign speakers whose languages were incomprehensible to them (Jahoda & Krewer, 1997, p. 5). 100 However, as I noted in Chapter 1 following Ashwell (2016), it is plausible some slurs (e.g., ‘slut’) lack neutral counterparts. 101 Anderson and Lepore (2013b, p. 352) give a similar example designed to motivate a distinction between acts of slurring and slurring words. 119 vocal impersonation (as opposed to a term), and NC is not meant to extend to such vehicles for slurring.

However, this kind of lexical gerrymandering seems ad hoc. Although we should distinguish slurring acts and slurring words, NC theorists need to give a principled reason for denying that the above counterexamples are members of the latter category. They should not simply insist that the class of conventional slurring words is, by definition, restricted to words and phrases whose truth-conditional content is identical to that of their neutral counterparts.

Doing so would beg the question by illicitly presupposing NC.

Moreover, there are independent reasons for thinking that some of the above counterexamples are conventional slurring words. First, if they are not conventional slurring words, then competent speakers should be able to use them as a purely descriptive, neutral phrases. Yet their occurrences in (18)-(20) seem non-defeasibly racist, despite the fact that the speaker is explicitly trying to use give a neutral description. Imagine that (18) is uttered in reference to a person who is of East Asian ancestry, (19) in reference to a Native American, and

(20) in reference to an South Asian person:

(18) He’s slanty-eyed (#No offense intended – I’m merely describing his epicanthic folds).

(19) He’s a redskin (#No offense intended – I’m merely describing his skin tone).

(20) He’s a curry muncher (#No offense intended – I’m merely describing how he consumes curry).

120

The parenthetical claims in each of the above statements, whose defectiveness is indicated by the hash sign, would serve only to compound the derogation, rather than cancel it.102 Granted, it may be impolite to describe or make salient a person’s physical features, even when the features in question are not unflattering or defective in any way, yet describing someone’s physical features in a genuinely neutral way and using descriptive slurs like the ones mentioned above seem importantly different. As evidence of this, observe the contrast between the following slurs and genuinely neutral descriptions: ‘slanty-eyed’/‘person whose eyes have epicanthic folds’; ‘curry muncher’/‘person who eats curry vigorously’. This suggests that ‘slanty-eyed’ and ‘curry muncher’ are indeed conventional slurs, and thus counterexamples to NC.

Of course, the non-cancellability of a slur’s offensiveness does not in itself tell us a great deal about its truth-conditional content (Jeshion, 2013a, p. 322). I have not offered a positive account here, since I hold that there is no such thing as the truth-conditional content of slurs.

Rather, there are many such contents. The point here is simply that a speaker’s inability to cancel the derogation accomplished by ‘slanty-eyed’, ‘redskin’, and ‘curry muncher’ indicates that these terms are conventional slurs, as opposed to neutral descriptions.

5. Conclusion

NC seems plausible as a thesis about single word noun phrase slurs such as ‘chink’,

‘dago’, ‘kike’, and ‘nigger’. However, these make up only one variety of a large array of conventional slurring words. Counterexamples to NC include slurs that contain truth-conditional content about the target’s (alleged) diet (‘curry muncher’), attire (‘raghead’), and physical features (‘slanty-eyed’, ‘nappy-headed’, ‘redskin’), as well as slurs that encode other kinds of

102 See Williamson (2010, p. 57) for similar data on defective attempts at cancelling the offensiveness of slurs. 121 non-classifying content (‘Jewish American Princess’, ‘camel jockey’). Since slurs make up a heterogeneous class, a uniform account of their truth-conditions is unlikely to succeed.

122

Chapter 5: Derogation Without Words: On the Power of Non-Verbal Pejoratives

1. Introduction

Although a large literature on pejorative language has emerged recently (see Chapter 2), derogatory expressions make up a much larger category, and theorists working on pejoratives have, for the most part, neglected non-linguistic phenomena. This Chapter focuses on non-verbal pejoratives (NVPs), including expressive behaviors designed to insult or denigrate a target constituted by gestures, the use of a certain tone of voice, the wearing of certain personal accoutrements, such as clothing or makeup, or other non-verbal expressions. This Chapter undertakes the following aims: (i) provide a taxonomy of NVPs, (ii) specify a set of explanatory desiderata that a successful theory of NVPs should meet, and (iii) motivate an original theory of the derogatory power of a subset of NVPs, namely those whose effectiveness crucially relies on iconicity. Along the way, I will point out some ways in which iconic NVPs differ from wholly arbitrary NVPs whose derogatory power does not depend on any sort of resemblance, and natural threat signals in the animal kingdom. While I lack the space to provide an exhaustive theory, I hope to set the stage for a broader study of derogatory communication that takes into account non-verbal expressive behavior.

I will not assume that NVPs make up a unified class. I delineate the class of NVPs by pointing to paradigm examples, such as the “middle finger” and derogatory impersonations employed in “Blackface” minstrelsy, and I offer an account that explains the rhetorical power of

123 these sorts of vehicles for denigration.103 I argue that the sort of derogation that an NVP accomplishes depends crucially on two factors: (a) whether its design is due to natural selection or convention, and (b) the kind of expressive vehicle it is (e.g., iconic vs arbitrary). Natural

NVPs like the human sneer share many features with facial expressions found in the animal kingdom, such as the wolf’s snarl and primate threat signals designed by natural selection to indicate an individual’s potential for aggressive behavior (Cartmill & Maestripieri, 2012).

Expressive behaviors of this sort are not very cognitively demanding on users or recipients. In comparison, iconic gestures invite recipients to form mental images of what they represent

(McNeill, 2005). Successful communication with these gestures requires, inter alia, abilities such as imitation and mental simulation (Tomasello, 2008). Thus, they require more cognitive sophistication. However, iconic gestures may lose their imagery over time in a process

Tomasello (2008) calls the “drift to the arbitrary.” One example is the hand configuration that constitutes the middle finger. At one point, this hand configuration was an iconic vehicle for inviting an image of sexual violation, yet it is now largely (though not entirely) conventionalized and arbitrary.

2. Categories of Non-Verbal Expressive Behavior

Non-verbal signals may be vehicles for what Grice called “speaker meaning”

(alternatively, “non-natural meaning” or “meaningNN”) when performed with appropriate communicative intentions (Grice, 1989, p. 215). By overtly pinching one’s nose, one could meanNN that one’s interlocutor omits an unpleasant smell. However, as we will see, pejorative gestures (especially those that are iconic) often have an imagistic or experiential component that is crucial to the derogation they accomplish, and this is not the sort of thing that can be fully

103 Walton (1993) and Egan (2008) take a similar approach in theorizing about metaphor and idiom, respectively. 124 captured in terms of any sort of propositional speaker meaning (or other types of content associated with different parts of speech).

Additionally, the performance of a NVP may, but needn’t, constitute the performance of an illocutionary act. We can imagine someone performing the act of refusing by flipping someone off. As I shall argue, many NVPs are powerful because they prompt observers to form mental imagery of various sorts, yet this needn’t be an illocutionary act just as smiling at a recipient in order to get them to smile back needn’t be an illocutionary act (see also Chapter 3).

The four main categories of non-verbal expressive behavior are gestures, facial expressions, miming, and extra-body non-verbal communication. ‘Gesture’ may refer to a type of action constituted by movements of the head (excluding facial expressions),104 limbs, or a certain bodily orientation, or it may refer to a type of expressive vehicle used in the performance of such acts (Bar-On, 2015). As we shall see, the act/vehicle distinction cuts across all four categories of

NVP. The phrase ‘middle finger’ may refer to the action of extending one’s middle finger in order to denigrate someone, or to a type of expressive vehicle, namely the hand configuration used in performing such acts. Theorists standardly distinguish emblematic gestures from iconic gestures.105 Emblematic gestures, such as the “Ok” sign, are designed by convention and their form has an arbitrary relationship to what they represent, whereas iconic gestures represent things, in part, by bearing a certain perceptible similarity to them (McNeill, 2005).

Two paradigmatic pejorative gestures are the middle finger (see Figure 1) and bras d’honneur (alternatively, the “Italian salute” or “forearm jerk”), in which one bends either arm at

104 Gestures may crucially involve the face, as when one points in a certain direction with their gaze (Kendon, 2004, p. 310). However, theorists standardly distinguish facial expressions, which (as I explain below) are typically understood as mere natural affective displays on the face, from gestures, which are used to refer to things. 105 Two other categories that I will not discuss in detail here are deictic gestures, which constitutively involve the use of the head, limbs, or a certain bodily orientation in securing a referent (e.g., pointing), and beat gestures/gesticulations, which are movements designed to highlight accompanying speech. 125 the elbow, raises it, and makes a fist with the bent arm while simultaneously gripping the bicep of the bent arm with the opposite hand (see Figure 2) (Morris et al., 1979, p. 80). Some pejorative gestures target specific racial groups. Consider a racist gesture in which one pulls the skin near one’s eyes sideways, thereby giving them a “slanted” appearance, in order to denigrate

East Asians. While the middle finger and bras d’honneur are not conventional vehicles for slurring particular groups, performances of these gestures could betray a person’s bigotry. In flipping off an Italian, one may intentionally and overtly manifest prejudice toward Italians, yet this gesture is not itself a conventional vehicle for denigrating Italians.

Figure 1:106 Figure 2:107

Gestures may have interesting rhetorical effects even when not performed overtly.

Following Green (2007, p. 66), an action ψ is performed overtly iff one ψs “intending that (a) something be publicly discernible, and (b) this intention itself be publicly discernible as well.”

106 Copyright Ralph DiFranco, 2016. 107 Copyright Ralph DiFranco, 2016. 126

Given this understanding of overtness, non-overt performances of NVPs are possible. Kendon, reporting on findings by Morris et al. (1979), claims that in Malta one can be prosecuted for publicly performing the bras d’honneur. As a result,

the Maltese have evolved a way of performing this gesture so that it could be mistaken for a mere rubbing of the arm, and not a gesture at all. In this version the left arm is held straight with the hand clenched in a fist, while the right hand gently rubs the inside of the left elbow. (Kendon, 2004, p. 10)

Apparently, the Maltese sometimes perform the gesture in a non-overt way so that it could pass for non-expressive behavior.

Pejorative facial expressions make up a second category of NVP. Facial expressions are typically understood as configurations of an organism’s face that signal the presence of an emotional state in the organism (see, e.g., Ekman, Davidson, & Friesen 1990; Green, 2007).

Some facial expressions are a result of ritualization at the phylogenetic level. At some point wolves began to forego biting during confrontations with other organisms in favor of displaying their teeth, which is the initial stage of a bite. Wolves who did so gained an adaptive advantage, as did recipients who withdrew upon seeing teeth, and over time this snarling behavior became genetically fixed (Tomasello, 2008, p. 22). As a result, wolves signal their aggressiveness by snarling.

A third category, miming, is the use of one’s body or voice in depicting something.108

Miming always involves (constitutively) an iconic representation of a target, often by way of a depiction or portrayal of an agent or group in which a performer reproduces certain (purported) characteristic features of the agent or group in question. Pejorative miming may involve

108 Wilhelm Wundt used “mimic gestures” to refer to gestures that directly imitate some object or action (Kendon 2004, p. 91). Similarly, Tomasello (2008) uses “pantomiming” to refer to gestures that visually depict things. As I explain below, miming is a much broader category, since members of this class needn’t involve a visual element. 127

(constitutively) costumes and makeup designed to depict a target in an unflattering way, yet it may be entirely non-visual, as when one performs a caricatured vocal impersonation of an individual or group. One could exaggerate idiosyncratic vocal features like a lisp, a monotone delivery, or the nasal quality of an individual’s voice.

Miming may disparage particular groups. In Blackface minstrelsy, performers (who were, for the most part, white) would darken their skin using burnt cork or face paint, wear wigs, dress in oversized, ragged costumes, sing songs, and perform dances and skits that depicted African

Americans as characteristically unintelligent, lazy, buffoonish, and cowardly. Visual imagery was crucial to these depictions. Blackface makeup invited an image of African Americans that depicted them as having unusual facial features (e.g., extremely large lips and eyes). Auditory imagery was also crucial. Performers spoke in a grammatically sloppy dialect, which suggested that their targets were inept speakers. The earliest Blackface minstrel shows in the US were performed in the 1830s, and at least one tradition survived in the Adirondacks until the 1950s

(Brown, 2013, p. 92). Examples of more recent instances of pejorative miming include

Hollywood’s depiction of East Asians and Native Americans, known as “Yellowface” and

“Redface,” respectively (Mag Uidhir, 2013). Particularly egregious examples of the former include Marlon Brando as Sakini in The Teahouse of the August Moon (Daniel Mann, 1956), and

Mickey Rooney as Mr. Yunioshi in Breakfast at Tiffany’s (Blake Edwards, 1961). Like

Blackface, these portrayals typically employed makeup and prosthetics designed to alter a

(usually white) actor’s appearance, as well as a caricatured accent.109 In some cases, the

109 As Mag Uidhir (2013) observes, a race mismatch between an actor and her character is not sufficient for making the actor’s portrayal offensive. Race mismatches widely considered to be innocuous include Eli Wallach as Tuco Ramirez in The Good, the Bad, and the Ugly (Sergio Leone, 1966), and Jenette Goldstein as Pvt. Vasquez in Aliens (James Cameron, 1986) (Mag Uidhir, 2013, p. 55). An actor- character race mismatch is also not a necessary feature of a derogatory portrayal. An actor who is a member of race R could denigrate her own group by playing a role that depicts Rs in a negative light. 128 responsible parties (including actors and filmmakers) may not have had an intention to denigrate anyone. They may have simply intended to produce entertaining works of fiction, yet this does not make their depictions of African Americans, East Asians, and Native Americans any less derogatory.

Of course, some cases of pejorative miming are intended to mock and ridicule targets.

The Guardian reports that during a 2014 football match in Spain between Levante U.D. and

Atletico Madrid, Atletico supporters mimicked primate gestures and vocalizations in an attempt to taunt Levante player, Papakouli Diop, who is Senegalese.110 In a different incident, the

Guardian reports that during a 2014 football match in Spain an opposing fan attempted to denigrate Dani Alves, who plays professionally for FC Barcelona, by throwing a banana at him.111 The crucial point is that derogation attaches to performances, which may or may not be motivated by an intention to denigrate. However, in general, miming requires, inter alia, an intention to invite one’s audience to form a certain mental image of the target (I return to this issue in Section 3).

The fourth type of non-verbal expressive behavior is extra-body communication between agents. One might describe sending flowers to someone as a “polite gesture,” and an action aimed at forestalling a conflict a “conciliatory gesture.” These are not literal gestures, but acts of communication constituted by something other than an act of speech or bodily movement.

“Racist gestures” not constituted by an utterance, a head or limb movement, a facial expression, or miming are also possible. One example occurred in 2014 on the University of Mississippi campus, when someone placed a noose on a statue of James Meredith, the first Black student

110 111 129 admitted to the University.112 Although the perpetrator(s) utilized only external props, it is nevertheless a form of derogation, and so a fourth category is needed.

A successful theory of NVPs should meet several explanatory desiderata. First, an account of the derogatory power of NVPs is needed. Several different kinds of derogation are possible. First, being targeted by, or observing the use of, a NVP may result in an immediate, visceral reaction: one may feel shocked or offended as a result of observing the performance of a prohibited gesture (cf. Anderson & Lepore 2013a on the offensiveness of slurring words).113

Similarly, a wolf’s snarl is designed by natural selection to intimidate recipients and get them to retreat by triggering a visceral, affective response (Tomasello, 2008). Derogation of this sort requires little by way of interpretation, and it needn’t leave a lasting impression on targets or bystanders. By contrast, some non-verbal signals are vehicles for a different kind of derogation that leaves a more permanent mark on targets by “sullying” them with unpleasant imagery. Most historians agree that Blackface minstrelsy promulgated prejudice toward African Americans

(Lott, 1993). Audiences may have found it difficult to quarantine the visual image of the buffoonish grins and burnt-cork covered faces, and the auditory image of vocal impressions that were characteristic of minstrelsy’s portrayal of African Americans (Toll, 1974). This sort of derogation is distinct from the visceral reaction of feeling offended or threatened, though both kinds of derogation may occur simultaneously.

Giovanni Bonaficio and Etienne Bonnot de Condillac claimed in the seventeenth and eighteenth centuries, respectively, that gestures have a power to convey content that is not easily captured with words (Kendon, 2004). In general, attempts to gloss what an agent communicated

112 113 As I noted in Chapter 1, we can make several fine-grained distinctions here. Offense could consist of the feeling that one has been personally slighted or it might consist of being upset at having observed something one takes to be obscene. One needn’t feel personally insulted in the latter case. 130 with an NVP of the form ‘She said/meant that p’ tend to be unsatisfactory, especially those that crucially involve a portrayal.114 This is not to suggest that it is always impossible to articulate what a speaker means by performing a certain gesture or facial expression. With a sardonic smile accompanied by a monotone ‘ha ha’, one could meanNN that one is not amused (Wharton, 2009, p. 113). However, NVPs that constitutively involve a portrayal often have a sensory or perceptible component that is not verbalizable, yet crucial to their derogatory power. When Mary mocks George by speaking in a shrill tone of voice that purports to mimic George’s, the qualitative features of Mary’s vocal impersonation play a crucial role in the derogation she accomplishes. Seemingly, saying or suggesting that an individual’s voice is harsh and irritating is importantly different from embodying these qualities in a vocal performance that supposedly mimics the target.115 A successful theory of NVPs needs to explain (or explain away) this difference, and more generally, the apparent non-translatability of NVPs.

Finally, a successful theory should explain the role of iconic representation in NVPs.

According to a longstanding view that dates back to Plato, iconic resemblance is neither necessary nor sufficient for representation (Dipert, 1996). In the Cratylus, Socrates argues that a word’s resembling a referent r is neither necessary nor sufficient to make the word a name for r.

As Goodman (1968) notes, resemblance is both reflexive and symmetrical, yet representation is neither. It is trivial that everything resembles itself, but most things do not represent themselves, and typically, when an object A represents an object B, B does not represent A. Thus, if iconicity

114 Similarly, Potts (2007) claims that verbal pejoratives exhibit “descriptive ineffability,” i.e., competent speakers cannot provide an adequate gloss of their meaning using words. Potts takes this to “suggest that expressive content is not propositional, that it is distinct from the meanings we typically assign to sentences” (2007, p. 177). See also Jay (2000). 115 Of course, a vivid description of a person’s voice may invite a hearer to form an auditory image of the target. The point here is that an imitation does this more directly in that it does not need to go through the medium of semantic processing in order to place constraints on the kind of image the recipient is to entertain. 131 is to play any important role in representation, it cannot be understood as mere resemblance. This raises the question: what features other than resemblance must an icon have in order to be a representation? Additionally, we need an explanation for how minstrel performances both portray African Americans, yet also promulgate an unflattering image of them. The fact that

Blackface is a type of mimesis raises the more general question: how can a portrayal be highly distorting such that its product bears little resemblance to the representatum, yet it is still a representation?

To summarize, a successful theory of NVPs should account for the following features:

(i) The derogatory power of NVPs: their power to affect both targets and bystanders, including (i) their immediate, visceral effects, and (ii) their more long-lasting “sullying” effects.

(ii) Non-translatability: why we often find it difficult to articulate what NVPs communicate, and to capture their derogatory power in terms of propositional speaker meaning.

(iii) Iconicity: for iconic NVPs, an explanation is needed for how their iconicity contributes to their derogatory power, and how an icon may both depict a target group, yet also promulgate a distorted image of them.

3. Convention, Natural Design, Idiosyncrasy, and Iconicity

The derogatory power of a NVP may be due to convention, natural selection, or idiosyncratic (individual) design. Conventions account for the rhetorical power of some NVPs.

Conventional expressive behaviors exhibit several hallmarks. First, they are used with a degree of regularity, which is supported by norms governing their performance to which all or most members of a community conform (Green, 2007, p. 144). Such norms may vary among different cultures, though this is not a necessary feature of a conventional signal. In western cultures, overtly displaying one’s tongue is widely considered to be an insult, yet many Tibetans consider

132 displaying one’s tongue to be a polite, affiliative expression.116

A second hallmark of conventional expressive behavior is arbitrariness. An arbitrary gesture’s form has no significant relationship to its meaning, in comparison with iconic gestures, whose form resembles the representatum in a certain way (McNeill, 2005, p. 48). The “Ok” gesture (in which one touches the tip of the index finger to the tip of the thumb, while the other fingers are spread out and point upwards – see Figure 3) is arbitrary in that nothing about its form makes it suitable for expressing approbation (McNeill, 2005, p. 48).

Figure 3:117

Further, the form and meaning of a conventional signal is pre-specified in that it does not admit of contextual variation (McNeill, 2005, p. 49). Although one may pragmatically express a great deal with a performance of the “OK” gesture, and the target of one’s approval is dependent on

116 According to a Tibetan legend, a malevolent ninth-century Emperor, Lang Darma, had a black tongue, and Tibetans began displaying their tongues as a way of showing that they are not black like Lang Darma’s (Shakabpa, 2010, p. 164). 117 Copyright Ralph DiFranco, 2016. 133 the context in which it is performed, the gesture’s approbation meaning is not contextually determined (this, of course, excludes ironic uses).

Some expressive behaviors are non-conventional in that their design is due to evolution by natural selection. Ritualized expressions like the wolf’s snarl needn’t be performed with anything like a Gricean communicative intention in order to signal aggression (so, wolves needn’t snarl with an intention to get a recipient to withdraw by means of the recognition of that intention). Some facial expressions are ritualizations of unpleasant behavior, such as the act of expelling something noxious. We can imagine a party guest wrinkling her nose, raising her upper lip, and then dry-heaving in response to the host playing loud music that she strongly dislikes.

One intermediate case between conventional and natural gestures is bowing, which naturally embodies deference or submissiveness, yet it may be somewhat conventionalized. Other intermediate signals may include grammaticalized expressions like ‘Yuk!’ and stylized imitations such as ‘ha ha’ (Wharton, 2009, pp. 99-100; see also Bar-On, 2015). Thus, non-verbal expressive vehicles fall on a continuum:118

Continuum 1: from vehicles designed by convention to those designed by nature

Conventional design------Natural design (“Ok” sign) (bow; ‘Yuk!’) (sneer)

Seemingly, there are no pejorative human gestures whose design is due to natural selection, though such gestures could have evolved through ritualization had our ancestors developed the habit of attacking with the limbs in a specific pattern, and the movements that constituted the attack became segmented and ritualized phylogenetically into a signal of an impending attack.

Some non-conventional expressive behaviors are due to individual design, as opposed to

118 The idea that gestures make up a continuum from the conventional to the natural can be traced to nineteenth century theorist, Andrea de Jorio (Kendon, 2004, p. 49). 134 cultural design or natural selection (Tomasello 2008, p. 21). In George Eliot’s Middlemarch, the character Rosamond Vincy cranes her neck whenever she is feeling obstinate. According to

Green (2007), such behavior could express Rosamond’s feeling of obstinacy if she craned her neck overtly, intending for others to come to know what she is feeling and for them to recognize this very intention, and if enough people followed her lead, the neck-twisting behavior could eventually become a conventionalized gesture (Green, 2007, pp. 142-3). One could even improvise slurring gestures. Imagine someone overtly pinching their nose when around members of a certain race, as though members of the race in question emit an unpleasant odor. Miming makes up an intermediate category between expressive behaviors whose design is due to convention and those whose design is idiosyncratic, since there may be more or less routinized and conventionalized ways of miming various actions or events, yet variation is often allowed.

Thus, non-verbal expressive vehicles make up a continuum from the conventional to the idiosyncratic:

Continuum 2: from vehicles designed by convention to those designed by individuals

Conventional design------Idiosyncratic design (“Ok” sign) (various iconic gestures) (Rosamond’s neck craning)

As Tomasello (2008) observers, some pejorative gestures probably started out iconic, yet their connection to iconic imagery faded as they became conventionalized over time (I return to this point in Section 3). This suggests that the arbitrariness of gestures makes up a continuum

(McNeill, 2005, p. 10; Tomasello, 2008, p. 219). While the middle finger gesture exhibits regularity and normativity, it may be embellished so that the performer mimes the act of inserting the finger into an orifice. In that case, its derogatory power is due, in part, to its iconicity, and so the relationship between the gesture’s form and its meaning is not completely

135 arbitrary.119 Other forms of pejorative miming may be somewhat conventionalized. When

Blackface minstrelsy was widely performed, it was governed by culturally specified norms and stereotypes that determined how Black characters were to look, sound, and behave (Lott, 1993).

Members of the category I am calling ‘extra-body non-verbal communication’ may crucially involve iconic imagery, though they needn’t. The noose around the neck of the statue of James

Meredith contains an iconic element in that it resembles the act of lynching. Another example is the incident in which a football fan threw a banana at FC Barcelona player, Dani Alves, during a match in the way that one might expect food to be given to a non-human animal such as an ape

(someone in the stands had thrown the banana on the ground next to him). Of course, the invitation to form imagery in each case depends on a number of assumptions being mutually accepted by observers (e.g., in the case of Alves, that the banana toss was not merely a practical action aimed at providing Alves with a snack mid-match, that apes characteristically eat a lot of bananas or enjoy them, etc.). The perpetrator was presumably also relying on racist assumptions about Alves’s appearance and his Brazilian ancestry. As the above examples indicate, non-verbal expressive behaviors may be placed along a third continuum from the arbitrary to the iconic (see also Kendon, 2004, p. 106):

Continuum 3: from arbitrary vehicles to iconic vehicles

Arbitrary------Iconic (“Ok” sign) (middle finger) (vocal impersonations)

4. Imagery and the Rhetorical Power of NVPs

The derogatory power of many NVPs is due to imagery they invite recipients to entertain.

As we have seen (in Chapter 3), imagery can be seductive because of its contagion effects.

119 The gesture’s form is arbitrary in that there is nothing about the extended middle finger that makes it especially well suited for miming acts of sexual violation compared with any other digit. 136

Images of a targeted individual or group that we do not quarantine tend to stick with us, and are likely to affect how we think about them and treat them.

4.1. The Role of Imagery in Iconic Representation

An iconic gesture may be a means of coordinating behavior in pursuit of a joint goal.

Once observers understand what the gesture represents, they must do some additional inferential work to understand the performer’s social aim (Tomasello, 2008, p. 66). One may solicit help harvesting potatoes by performing an iconic gesture that depicts digging, but success requires that recipients comprehend the communicative intention behind the gesture (Tomasello, 2008, p.

203). However, an iconic gesture may just be a means of soliciting recipients to form an image of something, and in that case the cognitive load on recipients is somewhat lighter, since they needn’t infer any further communicative goal.

In general, a representation x is an icon of y iff

(i) there exists a set of observable features f1, f2, …, fn that x and y share, and

(ii) there exists an agent a who invites a recipient r to imagine x as y by 120 encouraging r to notice that x and y share f1, f2, …, fn.

Iconic gestures needn’t be done in order to be interpreted. By embodying an idea, image, or experience, a gesture may serve as an aid to practical reasoning or as a means of shaping one’s own thoughts. Thus, for gestures that satisfy the above definition, a may be identical to r, and further, successful gestures needn’t be performed overtly. An architect constructing a model may perform an iconic gesture not for an external audience, but simply to help herself visualize where a series of buildings will go and what they will look like. Broaders et al. (2007) report that

120 This account builds upon the definition of iconicity proposed by Dipert (1996), as well as the discussion of iconicity in Tomasello (2008, p. 68), who claims that iconic gestures “induce the recipient to imagine a corresponding real action.” 137 children who used gestures as an aid when explaining a solution to a math problem tended to better understand a subsequent lesson on related material (Cartmill, Beilock, & Goldin-Meadow,

2012, p. 135). One can easily imagine children gesturing in a similar way when working out solutions in isolation.

As we have seen (in Chapter 3), inviting an interlocutor to imagine something does not require being able to articulate or have any conscious awareness of what one has invited one’s recipient to imagine (Walton, 1990, p. 38). When a child playing with a model train invites a playmate to be the conductor, part of the intentional content of his request is his recipient’s state of mind, since he is inviting them to participate in a joint episode of make-believe, yet there is no reason to think that his intention to introduce an imaginative activity must be consciously introspectible. However, in general, taking something to be an iconic representation requires understanding the author as inviting an image, though this understanding may be implicit.

Further, some common ground may be required for uptake of an iconic depiction, including recognition of the features that one’s depiction shares with the representatum, yet this may be based on shared cultural knowledge, and needn’t be explicitly understood or acknowledged by authors or recipients (Tomasello, 2008, p. 79).

This account does not hold that mere perceptible similarity between two objects is sufficient for making either one an iconic representation of the other, and so it avoids Plato’s and

Goodman’s objections to resemblance accounts of iconicity. Another motivation for my account is that it allows for icons that are not artifacts.121 A non-artifactual icon is a naturally occurring phenomenon that does not resemble the representatum by design, yet it meets the two conditions given above. Consider Camelback Mountain, located in Phoenix, Arizona, which resembles a kneeling camel. Presumably, this was not due to anyone’s design, yet as long as there exists an

121 This is one respect in which my account of iconicity differs from that put forward by Dipert (1996). 138 agent who invites observers to imagine the mountain as a camel by means of getting them to recognize the shared resemblance, it is an icon.

The figure that standardly appears on a crosswalk sign is an iconic representation of a pedestrian crossing a street because it bears some minimal resemblance to the representatum, a pedestrian crossing a street, with respect to its shape, and its designer endowed it with that feature in order to invite pedestrians and drivers to imagine the figure as a pedestrian. Of course, the figure does not resemble an actual human body in several ways – it is two dimensional, it lacks hands and feet, it has no facial features, the head is not attached to the body, and so on.

However, the sign does not invite observers to imagine a two dimensional, faceless pedestrian with no hands, feet, or neck. Thus, I shall call the figure on the crosswalk sign a non-distorting iconic representation (NDIR). NDIRs do not invite observers to attribute features to the representatum that it does not actually have. If an icon invites observers to form an image of the representatum in which it is depicted as having features it does not actually have, I shall call it a distorting iconic representation (DIR).

The distinction of NDIRs from DIRs will help illuminate the derogatory features of the iconic representations involved in Blackface minstrelsy. The elements of African American culture that minstrel performers “borrowed” include speech patterns characteristic of African

American English (AAE).122 Actors were likely exposed to several African American dialects, including West African Pidgin English, Plantation Creole, and an early version of AAE, through dialogues featuring Black characters in plays and novels, as well as transcriptions of slave narratives in various forms of print media, including sympathetic representations in abolitionist publications that did not intentionally distort speech for comic effect (Mahar, 1985, p. 273). AAE has several lexical, syntactic, and phonological features that distinguish it from other dialects of

122 Also commonly referred to as African American Vernacular English (AAVE). 139

American English. One distinctive feature is the use of the aspectual marker be, which indicates a habitual condition, as in ‘That’s how they be’ (Green, 2002, p. 50). Also characteristic is the use of multiple negators in a single negative sentence, as in ‘I don’t feel no certain way about it’

(Green, 2002, p. 77). AAE also includes a consonant cluster reduction rule: the final consonant group in a word is reduced to a single consonant sound, e.g., ‘gift’ is pronounced as ‘gif’ (Green

2002, p. 109). When performing in minstrel shows, actors followed conventions of early AAE, and many audiences held that minstrelsy provided an authentic representation of African

Americans (Lott, 1993). However, Frederick Douglass accused performers of presenting a counterfeit image in order to “pander to the corrupt taste of their white fellow citizens” (Lott,

1993, p. 15). Zora Neale Hurston also objected that minstrelsy promulgated a misleading caricature of African Americans: “If we are to believe the majority of writers of Negro dialect and burnt-cork artists, Negro speech is a weird thing, full of ‘ams’ and ‘Ises’” (1997, p. 71). One syntactic feature alluded to here is use of the first-person singular form of ‘am’ in third-person

123 contexts.

Phonological features of AAE were often distorted in minstrel theatre and published dialogues. In “General Washington: a Christmas Story” (1900), the speech of African American characters is on the whole consistent with the rules of AAE, yet it contains /f/ in the /th/ environment (e.g., ‘through’ is pronounced as ‘froo’), which is not in line with AAE phonology and makes it seems as though characters have a speech impediment (Green, 2002, p. 175).

Additionally, certain phonological features of AAE were standardly exaggerated in minstrel performances for comic effect, e.g., a heavily drawled ‘Yaa-za’ (for ‘yes, sir’) would elicit a laugh from audiences (Green, 2002, Chapter 7).

123 Green (2002, p. 177) notes that in some published ex-slave narratives, ‘am’ is used in third-person contexts (e.g., “My mammy am on the Kilgore place…”), yet she holds that there is not enough evidence to determine whether this use of ‘am’ was a conventional feature of early AAE. 140

Furthermore, song lyrics and dialogue in skits performed during Blackface shows were often riddled with mispronunciations and malapropisms. In the minstrel song “Jump Jim Crow,”

‘scientific’ is deliberately mispronounced as ‘skyentific’.124 English actor Charles Mathews claimed that while viewing a production of Hamlet starring the Black Shakespearean actor Ira

Aldridge at the African Grove Theatre in New York City, he was inspired to do a minstrel version of the play. In Mathews’ version, Hamlet’s soliloquy in act three, scene one read as follows: “To be or not to be, dat is him question, whether him nobler in de mind to suffer or lift up him arms against a sea of hubble bubble and by opossum (oppose ’em) end ’em.”125

The vocal impersonations standardly employed in minstrelsy constitute iconic representations of African Americans because these performances invited audiences to imagine performers as African Americans (in part) by getting them to notice the similarities between the performers’ speech and that of actual AAE speakers. These impersonations were DIRs to the extent that performers invited audiences to imagine African American speech as exhibiting various features it did not actually have. These portrayals did not merely lead audiences to make a factual error about the rules of AAE; they depicted targets in an unflattering way. Given empirical data on contagion, it is likely that imaginatively engaging with minstrel performances that caricatured AAE in the abovementioned ways led audiences to hear African Americans as lazy, unintelligent, and inept speakers.

Many visual aspects of Blackface minstrelsy were distorting in unflattering ways (see

Figures 4 and 5 below). Performers would apply greasepaint (or burnt cork) designed to darken their skin, and red makeup around their lips to make them appear larger (or, as in Figure 4 below,

124 The lyrics for this song are reproduced in their entirety in Lott (1993, pp. 23-4). 125 Mathews’ dialogue was originally transcribed in Marshall and Stock (1958, p. 40), and is reproduced in Lott (1993, p. 45). Aldridge rejected Mathews’ story, claiming that he had never attempted the character of Hamlet at the Grove (Lott 1993, p. 46). 141 performers would leave space between the performer’s lips and surrounding skin in order to achieve the same effect). Mark Twain, a self-proclaimed fan of minstrelsy, described the lips of performers as “thickened and lengthened with bright red paint to such a degree that their mouths resembled slices cut in a ripe watermelon” (Griffin & Smith, 2013, p. 295). Other sources of imagery include facial expressions designed to exaggerate facial features such as wide eyes and a large grin. Performers also employed a stylized gait. Additionally, they would wear ragged costumes (see Figure 6). Costumes would sometimes feature enormous collars and shoes several sizes too big, which had the effect of infantilizing Black characters (Lott, 1993, p. 143).

Figure 4:126 Figure 5:127

126 This image of a photograph of minstrel performer, Billy Van, was archived in the Harvard Theatre Collection at the Houghton Library at Harvard University, and is accessible online at: . 127 This advertisement for a minstrel theatre show starring performer Billy Van was archived by U.S. Library of Congress and is accessible online at: . 142

Figure 6:128

Because portrayals in minstrelsy crucially involved various kinds of visual and auditory imagery

(in the costumes, makeup, vocal impersonations, choreography, and so on), we cannot understand the derogation they accomplished in terms of any sort of verbalizable meaning.

As noted in Chapter 3, attempts at quarantining imagined content regularly fail in predictable ways because our mental architecture often processes imagined content in the same way it processes content that we take to be reflective of reality, even when we know we are merely imagining or pretending (Gendler, 2006, pp. 183-4). The effects of a caricatured vocal imitation may be even more striking when combined with visual elements, as in Blackface minstrelsy and the racist “ape” imitation directed at Senegalese athlete, Papakouli Diop.129 One

128 The source of this image, , offers the following description: “Cover to early edition of Jump Jim Crow sheet music. Thomas D. Rice is pictured in his blackface role; he was performing at the Bowery Theatre (also known as the ‘American Theatre’) at the time.” 129 The latter case is a bit more complicated than typical straightforward impersonations. Correctly interpreting the fans’ gestures and vocalizations requires seeing them as an invitation to imagine a non- human primate, as well as an invitation to see certain purported similarities between Diop and non-human primates. In this way, the fans’ use of miming is metaphorical (more on this in Section 5 below). 143 may find it difficult to quarantine all of the racist images, and this may increase the likelihood that entertaining such imagery influences how one views the target.

4.2. Iconicity and the Drift to the Arbitrary

Iconic gestures may be preferable to conventional communication devices, such as a shared language, when interlocutors do not share such devices (Tomasello, 2008, pp. 220-1). It would be surprising that the middle finger has crossed so many geographical, cultural, and linguistic boundaries if it was not an iconic gesture at some point in its evolution. Morris et al.

(1979, p. 81) claim that the middle finger was a popular sexual violation gesture in ancient

Rome, where it was known as digitus impudicus (the indecent digit). A shift from iconic gestures to arbitrary vehicles of communication is motivated by the need for communicative efficiency.

Once arbitrary communicative conventions emerge, everyone can rely on everyone else understanding them, and this makes communication easier (Tomasello, 2008, p. 221). Tomasello

(2008) illustrates one way in which a “drift to the arbitrary” can happen with iconic gestures:

Perhaps a female of the genus Homo wishes to go digging for tubers. …[S]he pantomimes digging…in the direction in which tubers are normally found. The cavemates understand…that this digging gesture is intended to depict a real instrumental action of digging. It is possible that some of them might then learn this gesture from her, by role reversal imitation, thus creating a shared communicative device that is conventional in the sense of being shared…. Some individuals not familiar with digging, perhaps children, observe this “Let’s go digging” gesture, and for them the connection between the ritualized digging gesture and the act of digging for tubers is opaque…; they think it is just intended to initiate leaving generally. They might then imitatively learn the gesture to initiate leaving…so that the original iconic grounding of the gesture is now completely erased. (This is not unlike the way that some motivated linguistic forms, such as metaphors, become opaque [“dead metaphors”] across historical time as new learners are not exposed to the original motivation.) (pp. 222-3)

Some iconic derogatory gestures may have undergone a similar drift. When the middle finger was first introduced, users may have performed it by thrusting the middle finger upward,

144 and when executing the bras d’honneur, at least initially, performers may have configured either arm into an ‘L’ shape, thrusting the ‘L’-shaped arm upward, while adding a simultaneous downward motion with the opposite hand on the bicep of the raised arm. In the United Kingdom in the 1970s, the bras d’honneur played a dual role as both a conventional insult and a crude sexual compliment. In uses of the latter type, it was standardly performed by men in reference to women whom they found sexually appealing, and the gesture was designed to embody the kind of sexual act the men would like to perform with the target, with the arm representing an erect phallus (Morris et al., 1979, p. 82). Recipients may have complied by using themselves as a prop in an episode of imagining, though this is not necessary for interpreting the gesture. The gestures themselves only loosely resemble acts of sexual violation, and so observers would have needed to fill in certain details themselves. Thus, it is easy to see how imaginatively responding to these gestures is likely to produce a sullying effect. When forming the relevant image, one participates in one’s own derogation, though one may form the relevant image quickly and spontaneously

(this is not to suggest that those who use pejorative gestures are not morally responsible for their derogatory performances.).

As these iconic derogatory gestures became more commonplace over time they were segmented and abbreviated, and as a result, performances lost much of their dynamism. In most contemporary performances of the middle finger, one extends the finger while keeping the arm and hand stationary. Likewise, the bras d’honneur became abbreviated into a more static gesture in which one configures either arm into an ‘L’ shape and places the opposite hand on the bicep of the raised arm without adding any motion. As this happened, each gesture’s connection to iconic imagery faded. Both gestures are still prohibited in many cultures (for instance, the bras d’honneur is subject to both social and legal prohibitions in Malta), yet for the most part, they

145 are no longer vehicles for the kind of derogation of which they were once capable. In most contemporary uses, they are vehicles for a more visceral kind of derogation – they leave us feeling shocked and offended – and this does not entail inviting an image. Social prohibitions on a gesture often remain even after the original basis for finding it offensive has been forgotten.

Presumably, the basis for prohibiting the middle finger and bras d’honneur in the first place was that these gestures invited disturbing imagery. As long as prohibitions are in place, the fully static, arbitrary versions of these gestures needn’t encode any content whatsoever in order to shock or offend recipients.

Notice, however, that a drift to the arbitrary at the level of communicative vehicles needn’t prevent performers from restoring a gesture’s imagery at the level of acts. One could always inject dynamism back into these gestures through one’s performance by adding one or more upward thrusts during a performance of the middle finger or bras d’honneur to create an iconic representation of sexual violation. One motivation for a drift back to the iconic from the arbitrary at the level of acts may be a desire to negatively affect targets by inviting disturbing imagery, a goal not easily accomplished with arbitrary gestures. Dynamic icons appear to facilitate imaginative responses that are experientially richer in comparison with static icons because they give the imagination more to work with. If so, we would expect dynamic icons to facilitate more robust contagion effects. Of course, one may opt to perform the more static, arbitrary version of a gesture when one wants to elicit a more visceral reaction that does not depend on imagery, or when the amount of time one has to perform the gesture does not allow for an elaborate pantomime (e.g., when one is passing another driver on the freeway).

To summarize, acts of derogation fall on a continuum from the static (largely arbitrary) to the dynamic:

146

Continuum 4: from static/arbitrary performances to dynamic/iconic performances

Static------Dynamic (static middle finger) (pantomiming sexual violation)

The kind of derogation accomplished by performances of the finger and bras d’honneur depends, inter alia, on where they fall on this continuum.

5. Imagery and Embodiment in Metaphorical Gestures

NVPs and approbative gestures that are metaphorical often make use of imagery. In general, gestures are especially suited to the production of metaphor in that they can embody abstract notions in imagery (Cartmill, Beilock, & Goldin-Meadow, 2012, p. 132; see also Lakoff

& Johnson 1980). Many such gestures involve miming constitutively, and thus they have an iconic component. One may perform a palm up open hand (PUOH) gesture that depicts presenting something with the hands in order to create a conduit metaphor (McNeill 2005, p. 49).

With a PUOH, one can “extend gesture imagery beyond depictions of concrete entities” and thereby convey content that is not easily imageable, such as the notion of “presenting” an idea to someone (McNeill 2005, p. 46). Similarly, a drawing that depicts strangulation could be an effective visual metaphor for an authoritarian state’s enacting draconian policies designed to stifle criticism of the government. Another metaphorical gesture, bowing, embodies submission toward one’s interlocutor by putting the performer in a lower position relative to them

(Kinsbourne, 2006, p. 210). Also consider a gesture that mimes masturbation. If it is in common ground that masturbation is taken to be a self-indulgent activity, the gesture may be used metaphorically to convey that someone is wasting time on a personal indulgence. Alternatively, in a non-metaphorical uses, the gesture could simply be a means of inviting observers to form an image that they would rather not entertain.

147

One possible explanation for why the thumbs up is expressive of approbation and the thumbs down disapprobation in their contemporary usage is that these gestures are a kind of abstract pointing. As Lakoff & Johnson (1980) observe, since ‘up’ is metaphorically linked to positive affective and physical states (an upright posture with feeling energetic) and ‘down’ to corresponding negative states (a drooping posture with feeling exhausted or ill), ‘She’s at the peak of her career’ is suitable for expressing approbation and ‘I have a low opinion of him’ is suitable for expressing disapprobation (Lakoff & Johnson, 1980, p. 15).

However, metaphorical gestures may evolve and become conventionalized over time, in much the same way that iconic gestures tend to drift to the arbitrary. A widely held view is that the disapprobation meaning of the “thumbs down” and the approbation meaning of the “thumbs up” originated in ancient Rome, where Coliseum crowds would decide the fate of a fallen gladiator by giving thumbs up (a vote to spare the gladiator’s life) or thumbs down (as a vote to slay the gladiator).130 Morris et al. (1979) argue that this view is mistaken and that those who wanted a gladiator slain would extend their thumbs and mimic a stabbing motion by repeatedly thrusting the extended hand upward, and would hide their hands and arms if they wanted to spare the spare the gladiator. Subsequently, in certain regions of Western Europe the extended thumb became an iconic image of a phallus and the gesture had a sexual violation meaning. As of the

1970s, the iconic pejorative meanings had faded, and the gesture took on an approbative meaning, though Morris et al. suggest that the approbation meaning may have been present earlier (p. 191).131

130 Morris et al. claim that the mistaken view is due to lazy translations of descriptions of gladiatorial competitions from Latin to English. 131 Morris et al. report that a field study they conducted between 1975 and 1977, in which they surveyed 774 people spread out among several different Western European countries, revealed that 738 people assigned the thumbs up gesture an approbation meaning, while 36 (most of whom were located in 148

6. The Inadequacy of Standard Semantic and Pragmatic Theories

Some may be inclined to reject the approach developed above in favor of a more economical view that extends a traditional semantic or pragmatic theory of to cover NVPs. In what follows, I show that such approaches are untenable, and so an alternative account like the one sketched above is needed. One possible (yet unsatisfactory) theory of NVPs follows the suggestion of Kaplan (1999) that semantic frameworks may be extended to cover phenomena commonly regarded as falling outside of semantics. A theory of this sort may hold that NVPs are just a different way of semantically expressing exactly what one could express with a word or sentence. Generalizing from the semantic theory of slurring words developed by Hom and May

(2013), perhaps in giving a target t the finger, one speaker means that t ought to be subject to negative evaluation. Call this the Semantic Elucidation Theory (SET).

SET faces several problems. First, it would fail for some of the same reasons that a semantic account of interjections fails. On the latter view, one who says ‘Ow!’ speaker means something like ‘I’m experiencing a sudden, unexpected pain’ (Wharton, 2009). If so, ‘Ow! I’m experiencing a sudden, unexpected pain’ should seem repetitive, yet it does not (Wharton, 2009, p. 79).132 Similarly, on a truth-conditional theory, the middle finger encodes something like ‘t is worthy of negative evaluation’, where t is the target. This view predicts that a performance of the middle finger directed at t followed by an utterance of ‘t is worthy of negative evaluation’ will be repetitive, since ex hypothesi, the gesture and utterance are identical in content. However, they do not feel repetitive; a speaker who says, ‘You’re worthy of negative evaluation’ and then flips

Southern Sardinia and Northern Greece) assigned it a pejorative meaning, something akin to ‘up yours’ (p. 189). 132 Wharton draws on the work of Kaplan (1999), who holds that ‘Ow!’ encodes expressive content, which, according to Kaplan, is distinct from the descriptive content encoded by statements like ‘I’m feeling pain’. 149 off her addressee does not repeat herself.133 This worry generalizes to whatever truth-conditional content proponents of SET assign to the finger. Also, notice that if one assigns to the finger the same content as a pejorative phrase like ‘fuck off’, a different non-repetition worry arises. A speaker who yells ‘fuck off!’ after flipping off an addressee does not repeat herself.

Additionally, propositional structures seem inadequate to the task of capturing the iconic elements in non-verbal expressions (Wharton, 2009, p. 125). As we saw in Chapter 3, words can paint a picture (take, e.g., descriptive, metaphorical, and iconic slurring words), yet SET lacks the resources to explain how gestures do this. Seemingly, Blackface minstrel performances are not offensive because they encode a proposition that (or speaker mean something along the lines of) African Americans merit negative evaluation. The various kinds of imagery that minstrel portrayals invite by means of the vocal impersonations, face paint, and costumes (among other elements) are crucial to their derogatory power, and as noted in Chapter 3, imagery itself cannot plausibly be understood in terms of any propositional content.

However, one could hold that while the images promulgated by minstrelsy and other offensive depictions of racial groups in fiction are not propositional themselves, these portrayals license certain pernicious false beliefs for export from the world of the fiction. Following

Gendler (2000), Mag Uidhir (2013) notes that works of fiction often invite audiences to export certain fictional truths from a story. In an example due to Gendler (2000, p. 76), a historical drama set in France during the reign of Louis XIV invites the audience to export certain propositions about French fashion during the time period (e.g., what the popular hairstyles of the period were). According to Mag Uidhir (2013), film fictions that depict racial, ethnic, nationality,

133 Potts (2007) argues that a truth-conditional theory of expressive words fails in a similar way. He notes that repetition of ‘damn’ (in, e.g., ‘Damn, I left my damn keys in the damn car’) does not lead to redundancy, but rather strengthens the emotional content of an utterance, whereas repeatedly describing one’s own mental state (as in ‘I’m angry. I forgot my keys. I’m angry. They’re in the car. I’m angry’) does not have the same strengthening effect (Potts, 2007, p. 182). 150 or gender groups in unflattering ways license propositions about these group for export. Mickey

Rooney’s portrayal of Mr. Yunioshi in Breakfast at Tiffany’s, for instance, licenses propositions for export that may lead to racist beliefs about ethnically Japanese people (Mag Uidhir, 2013, p.

62). Thus, while derogatory portrayals constituted by, e.g., makeup and a caricatured accent, do not themselves express propositions, they invite us to believe false propositions about the groups depicted by them. For Mag Uidhir, it is morally wrong to mislead people about issues such as race, ethnicity, and gender, and so a work of fiction is morally defective to the extent that it does this (2013, p. 65).

By locating the influential power of Blackface (and other problematic portrayals in works of fiction) at the level of belief, Mag Uidhir over-intellectualizes the phenomenon at issue.

Several authors point out that imaginatively engaging with fiction often affects us in non- doxastic ways. Audiences often get the physiological effects characteristic of fear when watching a movie monster devour its victims even though they do not believe it exists and they know they are not in danger while sitting in the theatre (Walton, 1978). Effects of this sort are not due to the audience exporting propositions from the world of the fiction. Gendler (2008) points to a study by Rozin et al. (1986) in which subjects felt disgusted when asked to eat a perfectly good piece of fudge that they themselves had formed into the shape of dog feces, even when they were consciously entertaining the thought that the fudge is safe to eat. They also report that subjects were hesitant to put their mouths on a sterile, newly-purchased vomit-shaped piece of rubber, despite believing that it is not real vomit. In general, engaging in an imaginative exercise or episode of pretense often leads to attitudes and behaviors outside of the context of the exercise that are in tension with one’s beliefs.134 Given these experimental findings, it would seem

134 Gendler (2008) explains these effects by positing a new category of attitude, aliefs, which unlike beliefs are a-rational, automatic, affectively laden states. Her idea that when one is asked to consume 151 possible for minstrel portrayals to negatively affect audiences with visual and auditory imagery even if they do not affect audiences at the level of belief. It is possible to walk into a theatre in which a minstrel show is put on consciously and explicitly believing that African Americans are not unintelligent, buffoonish, lazy, and cowardly, and walk out without exporting any false propositions about African Americans, yet still be sullied by the portrayals one has observed.

In light of the problems with propositional accounts, a theory that explain NVPs’ derogatory power in terms of non-cognitive attitudes may seem more plausible. Hornsby (2001) holds that slurring words express the same content as pejorative gestures, which on her view, are expressive of affective states like contempt for targeted groups (see Chapter 2). However, more needs to be said about how pejorative gestures like the middle finger express contempt for their targets, given that Hornsby is committed to denying that in giving a target t the finger, one says that t is contemptible. We can see how natural facial expressions like the sneer signal aggressiveness, but the evolutionary story that explains the sneer’s power does not illuminate the derogatory power of NVPs closer to the conventional side of the continuum. Finally, while it is plausible that imagery is a non-cognitive phenomenon, an additional theory (of the sort given above) is needed to explain the role of iconic imagery in NVPs.

A third approach is an extension of the relevance theoretic account of non-verbal signals developed by Wharton (2009). Wharton understands the expressiveness of non-verbal signals in terms of constraints they place on the kinds of inferences observers are to draw about the mental state of the signaler. This approach draws on Sperber and Wilson (1986), who posit the notion of an explicature, i.e., the maximally relevant proposition communicated by an utterance.

feces-shaped fudge, one believes that it is safe to eat and not feces, yet one has a “belief discordant alief with the content: ‘dog-feces, disgusting, refuse-to-eat’” (2008, p. 641). I will not assume that Gendler’s theory of aliefs is correct or that it accounts for the cases described above – the point here is simply that imaginative episodes often affect us in ways that do not alter our beliefs. 152

Recovering an explicature requires forming an ad hoc concept. If an interlocutor says, ‘The neighbors have a new dog’ in an alarmed tone and with a terrified facial expression, paralinguistic features of the utterance place constraints on the ad hoc concept DOG* (a dangerous or mangy dog) that hearers are to form in order to recover the explicature, which in the present case is something to the effect that the speaker thinks that their neighbor’s new dog is dangerous (Wharton, 2009, p. 55). However, as Wharton observes, some expressions, such as

‘yuk!’, needn’t be accompanied by an utterance that has propositional content. He holds that interjections by themselves may encode procedural information about how to recover a propositional attitude, a feeling, or a sensation (Wharton, 2009, p. 102). ‘Wow!’, e.g., “might activate (or add an extra layer of activation to) a range of attitudinal descriptions associated with delight, surprise, excitement, etc.” (Wharton, 2009, p. 90).

An extended procedural approach may account for a wide range of NVPs, yet it too is inadequate. Of course, acts of slurring may crucially involve paralinguistic features of an utterance that inform hearers about a speaker’s attitudes. A speaker who utters a neutral classifying term (e.g., ‘Jew’ or ‘Italian’) in a contemptuous tone may provide information that aids hearers in drawing inferences about the speaker’s prejudice (cf. Jeshion, 2013b). Similarly, when Mary rolls her eyes in response to George’s remark or says ‘I’m so glad George is here’ in a contemptuous tone of voice, her facial expression and tone may help observers draw inferences about her dissenting attitude. Notice that Wharton’s central cases include the use of facial expressions, tones of voice, and grammaticalized utterances, all of which have some sort of natural expressiveness. When Mary sneers at George, we can infer that she has taken an aggressive stance toward George because the sneer is a natural sign of aggression. Thus, if

Wharton’s procedural account can be successfully extended to cover the middle finger, we need

153 to know what kinds of attitudes (if any) this gesture is expressive of, yet an evolutionary story does not do any work here, nor does it work for any other NVPs that rely on iconic imagery.

Further, it is not clear that a presentation of the middle finger reliably tells us anything about the performer’s attitudes. The arbitrary version of the finger may be offensive simply because it is prohibited (cf. Anderson & Lepore, 2013).

Another question is whether relevance theory extends to cover racist NVPs. Blakemore

(2015) argues that the use of a slur is offensive because it provides procedural information that hearers can use to draw an inference about the speaker’s bigoted attitudes. More specifically, a speaker’s choice to use a slur for a group G instead of a non-pejorative alternative is a signal of her prejudice toward Gs (Blakemore, 2015, p. 32; see also Bolinger, 2015). It may be possible to assimilate racist NVPs to Blakemore’s relevance-theoretic account of slurring words. However, it is not clear why being given instructions about what to infer about someone’s attitudes would lead to the kind of interpersonal “stickiness” that racist gestures tend to provoke. Suppose the recipient knows exactly what kinds of inferences she is to draw about the performer’s attitudes based on their racist gesture – it is difficult to see why she should be sullied as a result.

A related approach is an extension of the expression as showing theory (EST), developed by Green (2007). Green distinguishes three varieties of showing: (i) showing that something is so (e.g., showing that a deductive inference is valid by doing a formal proof); (ii) showing-α, where α is an object of perception (e.g., showing one’s teeth by opening one’s mouth, thereby enabling others to see them); and (iii) showing how an experience, mood, or emotion feels (e.g., showing someone how vanilla smells by giving them a vanilla-scented candle) (2007a, pp. 47-8).

On an extended version of EST, NVPs are derogatory in virtue of what they show an audience. Perhaps performances of the middle finger show that one is feeling contempt for

154 someone. In that case, we need an account of how the middle finger does this, and we are led back to some of the same worries raised against approaches considered above. Merely showing someone my middle finger by making it perceptible does nothing to show the recipient how I feel about them, so the notion of showing-α does no work here. Another possibility is that the middle finger shows us how sexual violation looks through its iconic shape. Yet most performances of the middle finger bear only a very minimal resemblance to actual acts of sexual violation. In general, icons do not do a very good job of showing us how things actually look, and it is often not their job to show how something might look (e.g., the pedestrian icon on a crosswalk sign does not have the function of showing us how humans would look if we were faceless two- dimensional creatures with disembodied heads and no hands or feet). Rather, as we have seen, the function of an icon is to invite a mental image of the representatum. More importantly, an extended EST approach fails for the same reasons that Wharton’s procedural approach fails.

Showing one’s derogatory attitudes by providing evidence for them does not have the derogatory power that NVPs characteristically have. A historian could show us how African Americans looked to bigots in the early twentieth century by showing us footage of a Blackface minstrel performance from the era, but she does not thereby invite us to view African Americans that way. Even intentionally and overtly signaling the presence of one’s own attitudes by means of showing-how or showing-that is itself a neutral activity (Bar-On, 2015). One may do this without putting forward one’s attitudes as appropriate or fitting, as when one intentionally provides evidence of their defective attitudes to a therapist, not as a means of promulgating them, but to seek treatment for them.

7. Conclusion

We have seen that NVPs, like verbal pejoratives, make up a heterogeneous class.

155

Threatening a recipient by means of a natural signal like the sneer, and interpreting such threat signals, are not very cognitively demanding tasks, since the sneer is merely a ritualized attack.

By contrast, derogation by means of iconic NVPs requires some cognitive sophistication. To get the proper uptake, performances of iconic NVPs need to be interpreted as invitations to form imagery, and inviting recipients to do that requires, inter alia, an (at least implicit) understanding of their imaginative capacities. However, as Walton (1990) reminds us, invitations to imagine may be issued and complied with quickly, spontaneously, and without much effort, especially when props are used. Imagery is contagious in that it has the striking power to influence how we think, feel, and behave. Thus, iconic NVPs are often highly effective vehicles for derogation.

Iconic expressive vehicles may become conventionalized over historical time in a process that Tomasello (2008) calls the “drift to the arbitrary.” Arbitrary NVPs that were once iconic may still be offensive, despite losing their imagery, if they are socially prohibited. However, mere offensiveness is cheap. Users can transform a static, arbitrary gesture like the middle finger into a dynamic pantomime, and thereby re-introduce previous imagery. Thus, a move back to the iconic from the arbitrary at the level of acts is always possible.

While I have focused primarily on iconic NVPs, there is much more to say about their non-iconic counterparts. Further work must be done to explain the rhetorical power of natural emotional displays on the face, the process by which idiosyncratic gestures become conventionalized over time, and the nature of the taboos surrounding conventional pejorative gestures. This Chapter should be taken as an initial attempt to theorize about a rather large subset of NVPs, and to expand the literature on pejoratives beyond the verbal.

156

Chapter 6: Appropriate Slurs

It is natural to think that derogating people with slurs is wrong in principle. For it seems that no one should be subjected to derogation because of their group membership. At least, this assumption appears safe when the target group in question is a racial or ethnic minority. But what about cases in which the speaker using a slur is a member of an oppressed minority, and the slur in question targets a powerful majority group? Plausibly, members of an indigenous group in the U.S. who derogate white Americans with slurs need not be doing anything morally objectionable. Slurs may also be vehicles for protesting groups like the Ku Klux Klan, or they may be vehicles for humor (‘arm chair enthusiast’ for philosophers may be one example).

Slurring in these cases may be morally valuable in ways that slurring historically oppressed groups is not – or so I shall argue.

As we have seen, recent literature on slurs is devoted to accounting for why slurs tend to provoke offense in listeners (see Saka, 2007; Jeshion, 2013b; Camp, 2013; Anderson & Lepore,

2013a; and Bolinger, 2015, among others). Theorists have had much less to say about ethical questions surrounding the use of slurs, and they tend not to allow for permissible derogation. As

I noted in Chapter 2, Hornsby (2001) argues that slurs are “useless” for non-bigots such as ourselves, insisting that “there is nothing that we want to say with them” (p. 129). We also saw that for Anderson and Lepore (2013a), slurs are defined as prohibited words. Anderson and

Lepore (2013b) argues that such prohibitions have moral significance. On their view, groups

157 have a right to determine how they are referred to, and if the use of a slur for a group violates this right, its use is morally impermissible:

It is widely noted that…groups have a right for their culture to be respected, and perhaps, supported. Names are often important aspects of a group’s culture, and so, it is reasonable to include the manner in which a group is referenced as a part of its right to self-determination generally. If this is correct, it is a short step from a right to determine whether the use of a name is permissible to one to determine whether its use is impermissible (Anderson & Lepore, 2013b, p. 351).

It is implausible that all groups have a right to be respected. Presumably, groups whose self- understanding is partially constituted by a commitment to white supremacy (e.g., the Ku Klux

Klan and Neo-Nazis) do not have a right for their culture to be respected (Barry, 2001, p. 258;

Killmister, 2012, p. 263). In some cases, considerations about the importance of preserving a group’s culture and identity may actually provide a basis for slurring other groups. People who are deaf, e.g., may derogate people whose hearing abilities are typical, and thereby establish and reinforce a group identity. Whether the group consisting of people who can hear constitutes a distinctive culture (in Anderson and Lepore’s sense), they do not obviously have the right to unilaterally determine how deaf people refer to them. Thus, the sorts of considerations that

Anderson and Lepore give in favor of deciding whether the use of a name for a group is impermissible do not settle important questions about the ethics of using slurs.

In some cases, prohibitions on slurs seem appropriate and justified. We can imagine someone arguing that it is permissible to dehumanize Middle Eastern soldiers with racial slurs on the grounds that one is fighting in a just war against them, and slurring the enemy makes defeating them easier. Yet this view is misguided, since slurs that target Middle Easterners are likely to promulgate objectionable attitudes, and undermine the social standing and self-respect of innocent people both during the conflict and after it has concluded. However, as we will see,

158 the use of slurs needn’t give rise to wrongs of this sort. The aim of this Chapter is to show that derogation with slurs can constitute a morally valuable activity, and thus derogatory uses of slurs are sometimes morally permissible. I will focus on three sets of cases. In the first, slurs function as vehicles for protesting groups that manifest bad-making features (such as white supremacist groups). In slurring these groups as a form of protest, victims can maintain their self-respect.

Additionally, third-party observers may use protest slurs as a way of distancing themselves from the target group, and thereby reaffirming their integrity, and as a way of fostering solidarity with victims. In the second set of cases, vulnerable minority groups use slurs for a powerful majority to distance themselves from the majority, and thereby cultivate a unique group identity. Finally, the third set of cases consists of slurs that are conventional vehicles for a humorous type of derogation.

I will not provide an exhaustive account that purports to determine for each slur whether its use is permissible. I will not motivate cases of appropriation (See Chapter 3 for my own account). Nor will I address uses of slurs that would normally be impermissible, yet may be acceptable in the context of satire or a work of fiction designed to lampoon or criticize attitudes promulgated by paradigmatic derogatory uses of these terms (cf. Richard, 2008, p. 12; Bolinger,

2015). Rather, I will motivate the claim that the practice of derogating certain groups with slurs plays a positive role in users’ moral lives in a range of cases.

1. Slurs and the Varieties of Social Groups

While this Chapter focuses on the ethics of using slurs, it will be helpful to keep in mind that a sharp distinction cannot be drawn between slurs and other pejoratives. As we saw in

Chapter 1, some theorists distinguish slurs from pejoratives that target people on the basis of personal traits or behavior (Saka, 2007, p. 148; see also Hay, 2013, and Nunberg,

159 forthcoming).135 Yet, as several theorists point out, many derogatory terms applied on the basis of both group-membership and behavior are widely held to be slurs (see, e.g., Jeshion, 2013b;

Ashwell, 2016). Religion-based slurs and slurs for sexual orientation, for instance, do not clearly fall on either side of the group/behavior dichotomy. Further, as I noted in Chapter 2, membership conditions that determine who a slur targets may be based on agency. Some slurs target people on the basis of nationality, and one may voluntarily emigrate or defect to a nation.

While theorists acknowledge the value of appropriation, previous philosophical work has not addressed the question of whether group derogation with slurs (broadly construed) has a positive role to play in our moral lives. If slurring makes a positive contribution to users’ moral lives in a range of cases and does not incur the sorts of wrongs that are standardly attributed to derogatory uses of slurs, then we have reason to believe that slurring is morally permissible in some cases, contra Anderson and Lepore.

Some groups manifest what I shall call ‘bad-making features’. Plausibly, using slurs to protest groups for their bad-making features is a morally valuable activity (however, as I argue below, derogation of a number of groups that do not manifest bad-making features is also a morally valuable activity). The class of features that are bad-making is large and diverse, and I will not provide necessary and sufficient conditions for a feature’s being bad-making here. Such features may include commitment to the pursuit of socially destructive joint goals and plans

(such as genocide or other heinous activities), or commitment to an objectionable ideology. Two paradigm examples of groups with both are the Ku Klux Klan (hereafter, the KKK) and the

Golden Dawn, a Greek political party that has embraced a form of nationalism reminiscent of the

Nazi ideology (Ellinas, 2013, p. 549). Following Ritchie (2015), membership in organized

135 As I explained in Chapter 1, one alleged difference is that slurs have neutral counterparts, whereas particularistic pejoratives do not (see, e.g., Hay, 2013, p. 456). However, several theorists have raised objections to this view (see, e.g., Glezakos, 2012; Ashwell, 2016). 160 groups such as the KKK requires, inter alia, an intention to cooperate in shared plans and actions with other members. By virtue of joining the KKK, members are committed to the pursuit of the group’s harmful goals, including terrorizing racial and ethnic minority groups and LGBT people

(among other vulnerable groups), as well as a pernicious ideology that includes white supremacy. Suppose we create a slur for KKK members; say we call them ‘kukkers’. Slurring this group is a way of protesting their goals and ideals. The use of a protest slur has moral value in a number of ways. The KKK’s intended victims may use a slur to declare that the way this group proposes to treat them is unjust, and thereby maintain their self-respect (cf. Boxill, 1976).

Protesting a group with slurs may also be a morally valuable activity for people who are not among the group’s intended victims. In general, protesting a group is a way of distancing or dissociating oneself from them (Hill, 1979). Third-party observers who are not victims of a white supremacist group may slur them as a way of distancing themselves from the group.136 For third- party protesters who think that white supremacy has no place in their society, slurring this group is a way of living up to their moral commitments, and thereby maintain their integrity.137

Additionally, victim-centered moral considerations can motivate the use of slurs by those not targeted by racist hate groups. Protesting a white supremacist or anti-immigrant group may be an act of solidarity with vulnerable people targeted by such groups (cf. Harvey, 1999, p. 72). Insofar as slurs are vehicles for protest, they have an advantage over thick terms such as ‘racist’ and

‘bigot’, which are generic criticisms of an individual’s problematic attitudes or behavior that do

136 Camp (2013, p. 338) claims that slurs function to distance speakers from targets. 137 Bell (2013) argues that responding to people who manifest certain vices with contempt is a good way to protest their failure to meet standards one cares about and to thereby maintain one’s own integrity. My argument here does not depend on the success of Bell’s. If she is right, contempt is one form of protest, yet protest need not be contemptuous (see, e.g., Hill, 1979). As we have seen, some theorists hold that slurs are conventional vehicles for expressing contempt (see, e.g., Hornsby, 2001; Richard, 2008), a number of theorists deny this (see, e.g., Anderson & Lepore, 2013a; Camp, 2013). In any event, my aim here is not to motivate having a certain reactive attitude such as contempt, but rather to defend the practice of slurring in a range of cases. 161 not illuminate what is objectionable about organized groups like the KKK and the Golden Dawn.

Further, as Blum (2002) observes, generic thick terms like ‘racist’ have been so widely overused

(e.g., in situations where people manifest varying degrees of racial insensitivity or subconscious biases of different sorts) that they have lost much of their power as vehicles for condemnation.

We need not assume that protest’s value depends on educating one’s target about their bad-making features or compelling them to change. On some views, maintaining one’s self- respect is a valuable activity even when one knows that doing so will not motivate a change in those who threaten one’s self-respect.138 Likewise, as we will see, maintaining one’s integrity and protesting in solidarity with victims are plausibly valuable independently of whether doing so persuades the target of one’s protest to change or leads to other good consequences.

Derogation of a group may be impermissible when its members have been coerced into joining. Suppose that a strip mining operation has begun to contaminate a local town’s water supply. The miners, who were innocently unaware of the environmental effects of strip mining, learn about these destructive effects, and on the basis of their discovery, decide to end the operation. However, their supervisor has hired armed guards to force them to continue.

Ordinarily, participation in an environmentally destructive enterprise is a bad-making feature of a group, yet using a slur for the miners (if there were such a term) to derogate them qua participants in a strip mining operation, may be impermissible because they have been coerced.

Plausibly, under these circumstances, we ought to respond with sympathy, rather than using a slur to protest this group.

The activity of protesting a group by slurring them can play a positive role in our moral lives even when the target group is not as thoroughly evil as the KKK. Suppose we were to coin

138 Bernard Boxill suggests that to protest the violation of one’s rights is not necessarily to argue that one has the rights in question, and that often “people protest when the time for argument and persuasion is past” (Boxill, 1976, p. 63). 162 a slur for a group consisting of people working in the Hollywood entertainment industry. The target group includes, inter alia, studio executives, directors, producers, and actors whose work perpetuates an industry in which the status quo is the production of films that are likely to be financially successful, yet are uninspired and lacking in racial, gender, or cultural diversity.

Suppose we introduce the phrase ‘Hollywood hacks’ as a slur for this group. This phrase is a slur because it targets a class of people on the basis of the industry in which they work. In this case, the target is an instance of what Ritchie (2015) calls an ‘unorganized group’. Membership in groups of this sort does not require an intention to pursue joint plans and perform joint actions with the rest of the group’s members. Although filmmaking requires an intention to collaborate with others, and proper subsets of the target group corresponding to ‘Hollywood hacks’ do join organizations, such as entertainment industry unions, this group is unorganized because it is not the case that an individual’s membership is constituted by an intention to pursue common goals and joint actions with all of the other members. By derogating the target group as ‘Hollywood hacks’, speakers protest the lack of innovation, creativity, and diversity in Hollywood films.

Those who have a moral objection to the lack of diversity in Hollywood can use this protest slur to maintain their integrity. Additionally, a speaker could use this slur in an act of solidarity with marginalized artists; in making manifest their view that the treatment of these artists is unacceptable, observers can develop a valuable relationship with the artists (cf. Harvey, 1999, p.

71).

The practice of protesting organized and unorganized groups raises a number of philosophical issues concerning collective responsibility. Seemingly, organized groups can be responsible for what they do (Smiley, 2010). But what about unorganized groups? A number of theorists allow that groups without a decision-making procedure may be collectively responsible

163 for what they do.139 In the cases discussed so far, slurring is valuable because of its role in maintaining an agent’s self-respect (if she is a victim of the group being slurred), and in maintaining an agent’s integrity as well as cultivating solidarity with victims (if she is not a victim of the group being slurred). In general, we have no reason to think that the value of protesting a group through the use of slurs requires specifying an ostensible target. Speakers can protest the group consisting of Hollywood insiders by derogating them as ‘Hollywood hacks’ without knowing how responsibility for the lack of diversity in Hollywood is distributed over various studio executives, producers, and so on. It would of course be desirable to specify a discrete set of responsible individuals and hold them accountable, but that may be difficult or impossible in the present case.

To this point we have considered examples designed to illuminate the moral value of slurring groups that manifest bad-making features.140 While it may seem obvious that the activity of protesting and condemning these groups has a positive role to play in our moral lives, it is natural to suppose that we should never derogate people on the basis of features that are not

“bad-making.” However, a number of theorists observe that derogation with slurs may function to construct a group identity (see, e.g., Carnaghi et al., 2011; Nunberg, forthcoming), and as I will argue, the moral value of cultivating a group identity through derogation with slurs does not

139 Smiley (2010) attributes to Tuomela (1989) the view that crowds and mobs may be collectively responsible for harm and destruction, despite the fact that members may not intend to perform group actions, as long as some of the members directly contribute to harm and others either facilitate these contributions or fail to prevent them. 140 Camp (2013, p. 338) and Jeshion (2013b, p. 237) allow that derogatory attitudes toward certain groups may be warranted (in the sense of fitting). Camp holds that when one adopts a slurring perspective, one takes the property g, which determines membership in a group G, to warrant a negative response. She claims that “if g really did produce a range of properties that deserved to be condemned, then the corresponding emotions could be warranted” (Camp 2013, p. 338). Jeshion (2013b, p. 237) suggests that “individuals who exploit women and children by selling their bodies to others for sex are, and are widely seen as, deserving of contempt for that exploitation.” To claim that certain derogatory attitudes are fitting is not to claim that derogation with slurs is morally permissible. Further, neither Camp nor Jeshion purport to give an account of what is morally valuable about derogation. 164 depend on the target group’s manifesting bad-making features. In cases where derogation is valuable in virtue of its role in constructing a group identity, the target group is often not problematic qua group, and the slurs in question needn’t function as vehicles for protesting their targets. To see what is valuable about derogation of this sort, we will need to examine the ways in which a group may construct an identity by means of slurring another.

2. Relational Goods and Group Identities

In discussing the moral significance of group membership, Killmister (2012) asks readers to imagine “a minority group whose self-understanding was tied to a belief in the inferiority of white males…and whose shared practices involved…eulogizing about their inadequacies” (p.

265). According to Killmister, since “white males have consistent social reinforcement of their privileged position in society,” a marginalized group proclaiming them to be inferior is unlikely to negatively affect the group’s social standing or undermine members’ self-respect (2012, p.

265). Although Killmister does not mention slurs specifically, her suggestion provides a basis for a defense of using racial slurs (as well as other slurs that target people on the basis of membership in groups that do not exhibit bad-making features qua groups). However, successfully defending the use of slurs on the basis of Killmister’s suggestion requires explaining what is morally valuable about slurring for the sorts of groups she describes. If this approach is to succeed, the value of using these slurs must not be undermined by whatever wrongs (if any) speakers may incur in derogating targets.

We can imagine members of an indigenous sovereign nation within the U.S. derogating white Americans with slurs that target them on the basis of their nationality and race. The fact that the indigenous group is vulnerable and attempting to avoid assimilation, are morally relevant. By using slurs for the majority, and thereby distancing themselves from the target

165 group, indigenous group members construct a self-conception as members of a distinct autonomous group that has a unique culture. Derogation of the target with a slur is a powerful way of saying, in effect, ‘We’re still here!’ Given that preventing assimilation and maintaining a distinct group identity are morally worthwhile, we have reason to think that derogation of the majority is a morally valuable activity (I discuss the moral significance of group identities in greater detail in Section 3). We can even imagine a situation in which an indigenous language is dying, and one of the few remaining words is a slur for white people. To preserve the language, and prevent it from becoming a mere historical curiosity, speakers may need to use the slur.

Further, the use of a slur for the majority may be instrumentally valuable by fostering camaraderie among members of the indigenous nation,141 though its value as a vehicle for constructing a group identity needn’t depend on its use having this effect.

This is not to suggest that a group’s vulnerability is, by itself, sufficient for making its members’ use of a slur for a non-vulnerable group morally permissible. Plausibly, vulnerable minorities should not use slurs to establish a group identity based solely on differences with the majority, i.e., identifying themselves simply as “the Other,” which is the sort of approach to cultivating an identity that Simone de Beauvoir (1984) cautions against. One example of a potentially problematic slur is ‘breeder’, which in some LGBT communities is customarily used to slur heterosexuals (Queen, 2007, p. 319). The use of this term could be connected to cultivating a group identity. If the use of ‘breeder’ cultivates a self-conception on which speakers define themselves simply in terms of how they are presumed to differ from heterosexuals, then it may foster an unhealthy identity. One alleged difference that the slur hints at is that only heterosexual couples reproduce. Since the slur derogates heterosexuals for having children,

141 Relatedly, Lampert and Ervin-Tripp (1998) report that ridiculing members of an out-group serves to foster solidarity among the sexes. 166 perhaps its use constructs an identity on which LGBT speakers view themselves as people who do not (or should not) reproduce, and this may create a perverse disincentive for LGBT couples to have children.142 In that case, derogation with this slur may be impermissible.

In general, group identities will need to be assessed on a case-by-case basis. In many

Deaf143 communities in which members use American Sign Language (ASL), a conventional slur exists for those whose hearing abilities are typical. Deaf signers derogate hearing people with a pejorative version of a classifying sign called ‘ORAL’, which includes an iconic representation of “an exaggerated, large set of lips flapping” (Grushkin, 2003, p. 125; Padden & Markowicz,

1997, p. 424). This sign is a slur because its conventional role in Deaf communities is to derogate a class of people, namely hearing individuals who communicate with a spoken language, on the basis of their membership in that class. The ‘ORAL’ slur appears to be a vehicle for constructing and reinforcing a group identity. As with the case of indigenous groups slurring white

Americans, there needn’t be anything objectionable about identity construction in this case. In slurring hearing people, signers draw boundaries around their own tight-knit community and distance their community from the hearing world. The use of the slur appears to do this, in part, by highlighting one feature of Deaf culture that makes it special and valuable, namely the unique means of communication its members use. While a single, uniform Deaf identity that is stable across communities may not exist, a Deaf identity will often include a self-conception as a member of a small, cohesive group that uses a unique shared language, and whose members enjoy valuable shared experiences because of their shared features (Blume, 2010). It may also

142 We can also imagine antinatalists using ‘breeder’ as a slur for anyone who reproduces, rather than heterosexuals specifically. 143 Here I am following the convention, suggested by Padden and Humphries (1990), of using ‘Deaf’ to refer to a cultural group consisting of people who use a sign language and feel a shared sense of solidarity, and using ‘deaf’ to refer to individuals who have the audiological condition of being unable to hear, yet do not necessarily consider themselves members of a Deaf community. 167 include a disparaging attitude toward the use of spoken languages (Glickman, 1996, p. 124). Yet an identity of this sort is not obviously objectionable. Notice that ‘ORAL’ need not be a vehicle for protesting hearing people in the way that our imagined slur for the KKK is. By adopting the practice of derogating hearing people, and thereby distancing themselves from the hearing community, Deaf people make it the case that they are not simply people in a hearing community who cannot hear, but are instead people who constitute a separate cultural minority group.

Evidence suggests that the use of the ‘ORAL’ slur has not led to the wholesale exclusion of hearing people from Deaf communities. ASL also contains a laudative sign, ‘HEARING-

BUT’, which is conventionally applied to family members of Deaf people, interpreters who serve as mediators between deaf and hearing people, and others who are welcomed into the Deaf community because they exhibit respect for Deaf culture (Holcomb, 2013, p. 48). In addition to the moral value of maintaining a Deaf identity, use of the ‘ORAL’ slur may have instrumental value for its targets. For hearing people who are ignorant of Deaf culture, being the target of this slur could help them appreciate their insensitivity, and this may lead to a positive change in their attitudes and behavior, though the value of constructing a group identity need not depend on the slur’s use having this effect.

2.1. Humorous Derogation

Tim Jay observes that some pejoratives may be vehicles for humor or catharsis (2009, p.

89).144 Derogatory uses of slurs may also be permissible because they are vehicles for humor. In an episode of The Simpsons, the character Groundskeeper Willy denigrates the French as

‘cheese-eating surrender monkeys’. There needn’t be anything wrong with ridiculing the French on the basis of their (alleged) eating habits and military history (we can imagine an analogous

144 Jay’s discussion centers on swear words; he does not discuss the humorous slurs whose use I motivate in this section. 168 slur for U.S. Americans whose use seems acceptable: ‘hamburger-eating imperialists’).

Likewise, the use of mild slurs that target people based on their profession, e.g., ‘ivory tower dweller’ for academics, needn’t be objectionable.145 Similarly, there needn’t be anything wrong with an experimental philosopher playfully denigrating colleagues who work in certain other philosophical traditions by calling them ‘arm chair enthusiasts’. Another plausible case is

‘Shabbes goy’, a term that some Jewish speakers use for gentiles who help them do work when religious restrictions prohibit Jews from working. According to Orthodox laws, all work, including “lighting a fire,” is prohibited on the Sabbath, and a light bulb’s being on is considered a fire because it is constituted by a burning filament (Dundes, 2002). Thus, Orthodox Jews are not permitted to turn on lights during the Sabbath. Enter the Shabbes goy, usually a neighbor or friend, who provides a helpful service by switching lights on or off for Jewish people. ‘Shabbes goy’ does not target a specific group, and so it may be felicitously predicated of anyone who is not Jewish. This term appears to mark non-Jews as “the other,” yet it does so in a playful way, by gingerly ridiculing them as people who do menial tasks for Jews on the Sabbath, and moreover, the term is often just as much a vehicle for making light of the laws that restrict work as it is for mocking outsiders.146

These jocular slurs are not in the business of dehumanizing targets nor do they have the power to undermine their targets’ social standing in any other significant way (though this is, of course, an empirical question). While the amusement achieved in such cases often comes at the expense of the target, the kind of mild teasing accomplished with jocular slurs may be therapeutic for those who are slurred (cf. Boskin & Dorinson, 1985). The use of slurs like the

145 In certain political climates, ‘ivory tower dweller’ may not be so innocuous. If this slur is an effective means of discrediting higher education and motivating people to defund it, then derogatory uses of it may not be morally permissible, though I doubt this is the case. 146 I am indebted to Daniel Silvermint for helpful discussion on this issue. 169 ones mentioned above can encourage targets to view themselves in a less serious light and with a certain detachment, and may also get them to relinquish certain commitments they ought to rethink. Relatedly, being the target of the pejorative ‘ORAL’ sign could help a hearing person appreciate their own insensitivity to Deaf culture, and this may lead to a positive change in behavior. Notice that slurs for obnoxious groups, such as ‘Hollywood hack’, may have similar effects.

2.2. On the Significance of Vulnerability

Defending the use of slurs as vehicles for cultivating and reinforcing a group identity or as vehicles for humor requires motivating a distinction between privileged, vulnerable, and non- vulnerable groups. Groups have privilege relative to other groups. Heterosexual white men are a paradigm example of a privileged group, since they are privileged relative to most (if not all) other groups.147 Vulnerable groups are those whose status as proper objects of society’s protection and concern is threatened or undermined by widespread prejudice (Waldron, 2012, p.

5). Seemingly, the group consisting of people who can hear are not vulnerable (qua people who can hear). Jay (2009, p. 97) reports that vulnerable people are often the targets of the most harmful kinds of hate speech. For Waldron (2012), hate speech, which may constitutively involve the use of slurs, harms vulnerable minority groups by creating a social environment that undermines members’ implicit assurance that they can go about their daily business and freely participate in society in ways others can without being subjected to harmful discrimination.

Empirical data indicate that different slurs may promulgate different sorts of harm. The use of slurs for immigrant groups in their host nations predicts an increase in immigrant suicide rates (Mullen & Smyth, 2004). In contrast, slurs directed at non-vulnerable groups appear to lack

147 McIntosh (2000) articulates a number of privileges afforded to people on the basis of their perceived whiteness. 170 the power to undermine their targets’ social standing, and they do not foster the kind of oppressive social environment that slurs for vulnerable groups engender (Embrick & Henricks,

2013). To the extent that the use of a slur wrongfully undermines its target’s social standing (or leads to other forms of oppression), or cultivates an objectionable identity in users, its use is morally impermissible. Derogation with slurs may manifest both of these wrong-making features. In a study by Carnaghi et al. (2011), male subjects who self-identified as heterosexual were exposed to anti-gay slurs and were then asked a series of questions to assess their heterosexual identity (such as ‘I would feel nervous being in a group of homosexuals of my own sex’ and ‘I would feel comfortable with being labeled as homosexual’). The authors report that exposure to anti-gay slurs led subjects to adopt a traditional conception of masculinity and to distance themselves from homosexual men. They conclude that homophobic slurs function as vehicles for constructing and maintaining a collective heterosexual, masculine identity (Carnaghi et al. 2011, p. 1662). Thus, the use of anti-gay slurs may be objectionable not only because they undermine the standing of vulnerable groups in the eyes of the majority, but also because their use cultivates a problematic group identity. However, if slurs for non-vulnerable groups do not have these wrong-making features and they have a positive role to play in the moral lives of users, their use need not be objectionable (though, as I note below, borderline cases may be possible).

The fact that a group is vulnerable does not entail that its members are entitled to slur others or that the group’s members should not be slurred. Hate groups like the KKK and neo-

Nazis are vulnerable in that their views are widely rejected by the society at large, and their members are often ridiculed and ostracized because of their membership, yet it is plausible that derogating members of such groups with slurs as a means of reinforcing their vulnerability (qua

171 members of noxious hate groups) is permissible.148 The KKK is a paradigm example of a group whose vulnerability is warranted, in part, because of the reprehensible goals and plans it pursues, which include terrorizing members of vulnerable racial, ethnic, and sexual orientation groups

(among others). Further, these goals are motivated by white supremacy, anti-Semitism, and other objectionable attitudes that are central to the Klan identity (Ferber, 1998). White supremacists’ problematic identity helps explain why they should be powerless qua white supremacists. The reason that derogating people on the basis of their membership in a vulnerable or marginalized racial group is wrong is that no actual people should be disenfranchised because of their race.149

It is uncontroversial that the vulnerability of deaf people is not warranted. Sources of their vulnerability include institutional injustices such as limited access to opportunities for education due to a lack of adequate support services and facilities for deaf students at schools and universities (Thoutenhoofd, 2000, p. 263). Given that the KKK should be vulnerable because of features of its group identity, while Deaf and indigenous peoples’ vulnerability is not warranted and their group identities are not objectionable, we can see why it is permissible for the latter groups to use slurs for a powerful majority to reinforce their own group identities, yet impermissible for the former to do the same.

In some cases, the moral status of slurring in order to cultivate a group identity is less than obvious. The elderly are vulnerable in a number of ways, and their vulnerability is of course not warranted. It is not clear that the elderly have anything like a group identity, though such an identity could exist. Would it be permissible for elderly people to slur young adults (seemingly, a

148 I take it one can ridicule someone on the basis of membership in a white supremacist group without also disparaging them on the basis of membership in an innocent vulnerable group (e.g., being unemployed or living below the poverty line). 149 However, we can imagine beings that ought to be vulnerable because of the kind of beings they are. Suppose a “race” of parasitic alien invaders show up to Earth. Arguably, the use of slurs intended to make this group vulnerable need not be objectionable (see May, 2005, for a similar example).

172 non-vulnerable group) by calling them ‘snot noses’, e.g., in order to cultivate or reinforce a group identity? The answer depends, in part, on the kind of self-conception they would be nurturing in doing so, and this is not the sort of thing we can about know about a priori. Perhaps the use of ‘snot nose’ encourages some sort of anti-youth ageism among the elderly. If this is an objectionable way of thinking, then we may have reason to believe that the use of this slur by elderly people is impermissible. However, it is an open question whether all forms of ageism are in principle objectionable. Since we would need to know a great deal more about a group identity that includes ageism in order to make an assessment as to whether such an identity is objectionable, settling this issue is beyond the scope of this Chapter.

The defense of slurs offered here is intended to be neutral with respect to different normative theories. I have argued that slurring is valuable as a form of protest. A victim may slur a group to protest the way she has been treated by them, and thereby maintain her self-respect.

Third-party protesters may slur a group to maintain their integrity. For deontologists, self-respect and integrity are non-instrumentally valuable (Hill, 1979). On a consequentialist framework, however, protest with slurs may be valuable when it motivates members of a group to reflect on their bad-making features and to make a positive change (cf. Bell, 2013, p. 160). Additionally, slurring may also be a means of releasing built up tension in speakers suffering from oppression and injustice. Virtue theorists may think of the value of protest in a different way. In his

Nicomachean Ethics Aristotle suggests that failing to protest when one is wronged would make one foolish and slavish (1126a6-11). Finally, I have noted that the use of a protest slur by third- party observers may be morally valuable as an act of solidarity with those victimized by the target group. Solidarity with victims may be valuable on a consequentialist framework, e.g., when it alleviates the victims’ suffering. While the suffering of victims should matter to third-

173 party protesters, on some views the performance of an act of solidarity with victims is morally valuable even if the victims themselves are never aware of it, and even if no tangible good consequences result from it (see, e.g., Harvey, 1999, pp. 72-3).

One obvious consequentialist motivation for a vulnerable group’s reinforcing their group identity by means of slurring a powerful group is that doing so may lessen the burdens of vulnerability and marginalization (Heyes, 2012). However, as we will see, group identities may be necessary for establishing group rights. Some theorists argue that group rights are essential for group autonomy (see, e.g., Kymlicka, 1995; Killmister, 2012), and the value of autonomy can be acknowledged on a variety of different normative theories.

3. Pro Tanto Wrongs

One could object to the use of slurs by pointing to various pro tanto wrongs that speakers or signers may commit in derogating targets. First, non-vulnerable people may find the use of slurs for their group insulting.150 Suppose George, a white man whose hearing ability is typical, is deeply offended when he observes Deaf individuals using the ‘ORAL’ slur in reference to him, and when members of an indigenous group ridicule and disparage him for his whiteness.

Plausibly, the fact that a behavior is offensive provides defeasible evidence that one ought not do it, and so one may argue that the potential for offense undermines the positive role that slurring plays in the construction of Deaf and Indigenous identities. A different worry stems from the fact that one could be a member of a non-vulnerable group and a vulnerable group simultaneously.

This allows for individuals to be both privileged and oppressed in different ways (Young, 1990, p. 42). Given that privilege and vulnerability often overlap, it may seem difficult to determine

150 Croom (2015, p. 148) suggests that anti-white slurs can negatively affect whites by triggering stereotype threat, a phenomenon whereby one manifests negative stereotypes about one’s own group. However, it is not clear how widespread stereotypes about whites are. Further, Croom does not provide any evidence linking slurs to stereotype threat (either in whites or other groups). 174 who is a morally permissible target of slurs.

To further elucidate these two challenges and to see how best to answer them, it will be helpful to distinguish ostensible targets of a slur s, i.e., the individual(s) referred to with a slur on a particular occasion of use, from s’s target group (Jay, 2009, p. 83). The set of all hearing individuals constitutes the target group of the ‘ORAL’ slur in ASL. An individual may be a member of the group consisting of people who can hear, which is plausibly non-vulnerable, as well as a member of other groups that happen to be vulnerable. Suppose that Naomi, a child who is not a member of a Deaf community, has just received a cochlear implant, and so she is a potential ostensible target of the ‘ORAL’ slur. Even though Naomi is now privileged by virtue of being able to hear, she is also still vulnerable in various ways, and so we might worry that in this case derogation with the ‘ORAL’ slur may hurt Naomi more than someone whose hearing ability is typical.

In answering these challenges, we need not assume that the goal of reinforcing a Deaf or

Indigenous identity is morally significant in such a way that personal offense does not detract from its value. Fortunately, signers may perform the ‘ORAL’ slur without targeting any specific hearing individual, and only when no hearing individuals are present. In cases where a group’s identity-cultivating slur is a word, members may utter the term only within their own groups (or silently to themselves by means of inner speech). In general, we have no reason to think that cultivating a group identity with slurs crucially depends on the speaker or signer having any ostensible target, or on the presence of anyone who falls within the slur’s extension, or even on the target group’s knowing that they are being slurred. A group’s practice of using a slur may be

(and often is) an internal practice of the group. Thus, the use of a slur need not insult or offend anyone, and so worries about incurring offense need not arise.151 Appreciating the point that the

151 This is not to suggest that the use of slurs is wrong only to the extent that targets are offended. As I 175 use of a slur needn’t have an ostensible target helps deflate a different objection: slurring a powerful majority could lead to retaliation, and as a result, may endanger vulnerable people or their allies.152 Since targets needn’t be aware that they are being slurred, uses of slurs needn’t provoke retaliation.

Another concern is that ostensible targets may appropriate a slur for their own group in order to transform it into a positive label. This could potentially diminish the effectiveness of protest slurs. One effect of the KKK’s appropriating a slur for their group may be that the term can no longer serve as a way for speakers to distance themselves from the KKK’s group projects and ideals. For in that case, in calling members of the target group the appropriated slur, speakers would be using a term that whose conventional role is to show support for the target (in much the same way that ‘queer’ is now widely held to be a positive label for LGBT people). The possibility of target groups appropriating protest slurs poses a challenge only if we have reason to believe that successful protest with slurs requires that targets are aware that they are being slurred. As with the identity-construction cases, when the motivation to protest is to maintain one’s own integrity or when it is an act of solidarity with victims, we have no reason to think that protest requires that targets are aware of the existence of a protest slur. If the targets of protest are not aware that they are being slurred, reclamation of the slur is not possible. Further, a powerful majority group, e.g., hearing people, would not seem to have any motivation for appropriating a slur used only in small, isolated Deaf communities. Seemingly, the appropriation

noted above, the use of a slur may be impermissible because of its role in cultivating an objectionable identity, and utterances of a slur may do this independently of whether anyone finds them insulting. The use of anti-Semitic slurs, e.g., may be an internal practice of a small, secretive group that has little or no influence in the broader society. Given that an anti-Semitic identity is objectionable, and this group’s use of an anti-Semitic slur serves to cultivate such an identity, we have reason to believe that their use of the slur objectionable. 152 Cf. hooks (1995, p. 13) on the dangers of expressing rage as a member of a marginalized group: “We learned when we were very little that black people could die from feeling rage and expressing it to the wrong white folks. …Rage was reserved for life at home – for one another.” 176 of ‘queer’, for instance, was motivated by a desire to take away a powerful vehicle for ostracizing and persecuting LGBT people. Since use of the ‘ORAL’ slur in Deaf communities has not led to the marginalization of hearing people in the broader society, it is not clear what motivation hearing people would have to appropriate it.

A final worry about the use of identity-constructing slurs is that group identities are objectionable because they foster essentialist attitudes about group membership. One common concern is that a group identity will inevitably foster a homogenous conception of the group that leads to the exclusion of outsiders. If group identities are objectionable in principle, the use of slurs designed to cultivate group identities would be impermissible.

We needn’t assume that group identities are completely innocuous, even when the groups in question do not manifest bad-making features. Group identities are likely to arise, since strangers who encounter one another in new cities tend to “renew their ethnic, locale, age, sex and occupational group identifications” (Young, 1990, p. 47). A fortiori, a Deaf identity seems almost inevitable given the communicative differences among deaf people and the majority of hearing people. Further, one independent motivation for maintaining group identities is that they provide a basis for certain group-differentiated rights and entitlements (Kymlicka, 1995;

Killmister, 2012, 2014). These include the right to an interpreter, special claims to land, which

Aboriginal Australians are entitled to under the Native Title Act of 1993, exemption from local laws that place limitations on fishing and hunting, various forms of affirmative action, including preferential admission to universities, and protection from hate speech, among others (Killmister,

2014, p. 92). While some of the aforementioned rights may be controversial, the wholesale rejection of group-differentiated rights would be too quick.153 As Killmister (2014) observes, in

153 Killmister (2012) defends an account of group rights based on groups’ interest in preserving their identity and culture that takes into account challenges posed by noxious group (see, e.g., Barry 2001). 177 some contexts, members’ self-conception as members of a group may be an important factor in determining whether they have the relevant rights or entitlements. The Deaf community’s self- identification as an ethnic and linguistic minority provides a basis for their demand for resources to promote American Sign Language. The Bilingual Education act, e.g., “provides funding for a variety of programs promoting the use of minority languages in the schools, and civil rights statutes…impose an affirmative duty on the schools to give children who speak a minority language an equal educational opportunity by lowering the English language barriers” (Lane,

2005, p. 297). When Deaf people are understood as mere disabled members of a hearing society rather than an ethnic minority, they may lose these rights and entitlements (Lane, 2005, p. 206).

Adopting a pluralistic conception of group membership will help circumvent worries about essentialism, and will also help further alleviate concerns about intersectionality. Pluralism allows that in some contexts it may be sufficient for membership in an indigenous group that one lives in a community with indigenous people, speaks the language, and is entrusted with certain rituals. So, e.g., it is possible in some contexts for a child of Italian immigrants who lives in

Australia to be a member of an Australian Aboriginal group (Killmister, 2014). However, if the question is whether one’s group membership affords one special welfare benefits or preferential admission to universities, one’s lineage and whether they have suffered from a history of discrimination may be crucial in determining who counts as an indigenous person (Killmister,

2014). A pluralistic account of group membership allows that hearing people, including those who are hard-of-hearing, may be members of a Deaf community. It is plausible that hard-of- hearing people should not be excluded from Deaf communities, even though they may be classified as “hearing” in some contexts. Pluralism also allows that hard-of-hearing people should not count as “hearing” for the purposes of determining who is among the ostensible

178 targets of the ‘ORAL’ slur, since they often face challenges and discrimination by the society at large due to their limited hearing abilities.154 In fact, people who are not deaf are often welcomed into Deaf communities. Many hard-of-hearing people who had not previously conceived themselves as members of the Deaf community “have chosen to embrace a Deaf identity because of social and communicative difficulties in the Hearing world,” and such individuals who use sign language have no problem socializing within Deaf circles (Grushkin, 2003, p. 126).

Relatedly, Christiansen and Leigh (2004) report that deaf children given cochlear implants do not become isolated from their deaf peers and friends as a result. Further, as noted previously, translators and others whose hearing ability is typical may be accepted into Deaf communities under various circumstances. Thus, if the practice of slurring hearing people is widespread in

Deaf communities, it does not seem to predict the exclusion of people who are not deaf from these communities (though this is an empirical claim whose verification may require further observation). It seems, then, that the practice of constructing a group identity through derogation of hearing people need not be in tension with promoting group-diversity and rejecting an inflexible, essentialist group-identity.

4. Conclusion

Previous philosophical work on slurs has not addressed the question of whether derogation with slurs is a morally valuable activity. I have argued that derogatory uses of slurs have a valuable role to play in users’ moral lives in three sets of cases. In the first, slurs are vehicles for protest. The use of protest slurs is valuable for those who have been victimized by the target group (as a way for victims to maintain their self-respect) and non-victims (as a way of reaffirming their moral integrity, and as an act of solidarity with victims). Other slurs are

154 I will leave it open whether people whose hearing is typical and are not vulnerable in other ways are morally permissible ostensible targets of the ‘ORAL’ slur. 179 vehicles for constructing a group identity. One example is the ‘ORAL’ slur in ASL, whose function is to construct and reinforce a Deaf identity by distinguishing the in-group from the out- group as well as highlighting what is special and distinctive about Deaf communities vis-à-vis the hearing community. One motivation for maintaining group identities is that they provide a basis for certain group rights, which on some views are necessary for group autonomy. Finally, slurs may be vehicles for humor, which can lead to positive self-reflection.

In general, the value of slurring needs to be assessed on a case-by-case basis. There is a great deal more to say about when slurring groups is morally objectionable. At what point should one abandon the project of trying to change a group to which one belongs and resort to dissociating oneself from them with a slur as a form of protest? I take this Chapter to be one step in a series of larger, incremental projects. The step undertaken here was to illuminate the moral value of slurring in a range of cases. Future steps include determining how severe and incorrigible a group’s bad-making features need to be in order to make it an appropriate target of protest slurs, and determining the moral status of derogation in borderline cases in which a group’s identity is not clearly innocuous or objectionable.

180

References

Adams, F. & Campbell, K. (1999). Modality and abstract concepts. Behavioral and Brain Sciences, 22, 610. Allen, I. (1983). The Language of Ethnic Conflict: Social Organization and Lexical Culture. New York: Columbia University Press. Anderson, L. & Lepore, E. (2013a). Slurring Words. Nous, 47(1), 25-48. Anderson, L. & Lepore, E. (2013b). What Did you Call Me? Slurs as Prohibited Words. Analytic Philosophy, 54(3), 350-363. Aristotle (1999). Nicomachean Ethics, 2nd edition, translated by T. Irwin. Indianapolis: Hackett. Ashwell, L. (2016). Gendered Slurs. Social Theory and Practice, 42(2), 228-239. Ayer, A. J. (1936). Language, truth, and logic. London: Gollacz. Bach, Kent (1994). Conversational Impliciture. Mind & Language, 9(2), 124-162. Bach, Kent (1999). The Myth of Conventional Implicature. Linguistics and Philosophy, 22(3), 327-366. Bach, Kent (2014). Loaded Words. Unpublished manuscript. Retrieved from http://userwww.sfsu.edu/kbach/Bach.LoadedWords.APA.pdf Bar-On, D. (2015). Expression: Acts, Products, and Meaning. In S. Gross, N. Tebben, & M. Williams (Eds.), Meaning Without Representation: Expression, Truth, Normativity, and Naturalism (pp. 180-209). Oxford: Oxford University Press. Barsalou, L. (1999). Perceptual symbol systems. Behavioral and Brain Sciences, 22, 577-660. Barsalou, L. W., Simmons, W. K., Barbey, A. K., & Wilson, C. D. (2003). Grounding conceptual knowledge in modality-specific systems. Trends in Cognitive Sciences, 7(2), 84-91. Barsalou, L. & Wiemer-Hastings, K. (2005). Situating abstract concepts. In D. Pecher & R. Zwaan (Eds.), Grounding Cognition: The Role of Perception and Action in Memory, Language, and Thought (pp. 129-163). New York: Cambridge University Press. Barry, B. (2001). Culture and Equality. Cambridge, MA: Harvard University Press. Beauvoir, S. (1984). The Second Sex, translated by H. M. Pashley. Harmondsworth: Penguin. Bell, M. (2013), Hard Feelings: The Moral Psychology of Contempt. Oxford: Oxford University Press. Ben-Ze’ev, A. (2001). The Subtlety of Emotions. Cambridge, MA: MIT Press. Bianchi, C. (2014). Slurs and Appropriation: An echoic account. Journal of Pragmatics, 66, 35- 44. Blackburn, S. (2006). Antirealist Expressivism and Quasi-Realism. In S. Darwall, A. Gibbard, & P. Railton (Eds.), The Oxford Handbook of Ethical Theory (pp. 146-162). Oxford: Oxford University Press. Blum, L. (2002). ‘I’m Not Racist, But…’: The Moral Quandary of Race. Ithaca, NY: Cornell University Press.

181

Blume, S. (2010), The Artificial Ear: Cochlear Implants and the Culture of Deafness. New Brunswick, NJ: Rutgers University Press. Bolinger, R. J. (2015). The Pragmatics of Slurs. Nous, doi: 10.1111/nous.12090 Boskin, J. and Dorinson, J. (1985). Ethnic Humor: Subversion and Survival. American Quarterly, 37(1), 81-97. Borghi, A. M., Glenberg, A. M., & Kaschak, M. P. (2004). Putting words in perspective. Memory and Cognition, 32, 863-873. Boxill, B. (1976). Self-respect and protest. Philosophy and Public Affairs, 6, 58-69. Brandon, R. (2000). Articulating Reasons: An Introduction to Inferentialism. Cambridge, MA: Harvard University Press. Broaders, S. C., Cook, S. W., Mitchell, Z. & Goldin-Meadow, S. (2007). Making children gesture brings out implicit knowledge and leads to learning. Journal of Experimental Psychology, 134(4), 539-550. Brontsema, R. (2004). A Queer Revolution: Reconceptualizing the Debate Over Linguistic Reclamation. Colorado Research in Linguistics, 17, 1-17. Brown, L. (2013). Can American Popular Vocal Music Escape the Legacy of Blackface Minstrelsy? Journal of Aesthetics and Art Criticism, 71(1), 91-100. Camp, E. (2013). Slurring Perspectives. Analytic Philosophy, 54(3), 330-349. Camp, E. (forthcoming). Why Metaphors Make Good Insults: Perspectives, Presupposition, and Pragmatics. Philosophical Studies. Camp, E. & Reimer, M. (2006). Metaphor. In E. Lepore & B. Smith (Eds.), The Oxford Handbook of Philosophy of Language (pp. 845-863). Oxford: Oxford University Press. Card, C. (1995). Lesbian Choices. New York: Columbia University Press. Carnaghi, A. & Maass, A. (2007). In-Group and Out-Group Perspectives in the Use of Derogatory Group Labels: Gay versus Fag. Journal of Language and Social Psychology, 26, 142-156. Carnaghi, A., Maass, A., & Fasoli, F. (2011). Enhancing Masculinity by Slandering Homosexuals: The Role of Homophobic Epithets in Heterosexual Gender Identity. Personality and Social Psychology Bulletin, 37(12), 1655-1665. Cartmill, E., Beilock, S. & Goldin-Meadow, S. (2012). A word in the hand: action, gesture and mental representation in humans and non-human primates. Philosophical Transactions of the Royal Society B, 367, 129-143. Cartmill, E. & Maestripieri, D. (2012). Socio-Cognitive Specializations in Nonhuman Primates: Evidence from Gestural Communication. In J. Vonk & T. K. Shackelford (Eds.), The Oxford handbook of comparative evolutionary psychology. (pp. 166-193). Oxford: Oxford University Press. Carston, R. (2010). Metaphor: Ad Hoc Concepts, Literal Meaning and Mental Images. Proceedings of the Aristotelian Society, 110, 295-321. Chomsky, N. (1980). Rules and Interpretation. New York: Columbia University Press. Christiansen, J. B. and Leigh, I. W. (2004). Children with Cochlear Implants: Changing Parent and Deaf Community Perspectives. Archives of Otolaryngology: Head and Neck Surgery, 130(5), 673-677. Cohen, P. (2010). Discussing That Word That Prompts Either a Fist Pump or a Scowl. Retrieved March 13, 2016, from http://www.nytimes.com/2010/01/23/arts/television/23shore.html?pagewanted=all&_r= 0

182

Croom, A. (2011). Slurs. Language Sciences, 33, 343-358. Croom, A. (2013). How to do things with slurs: Studies in the way of derogatory words. Language and Communication, 33, 177-204. Croom, A. (2015). Slurs, Stereotypes, and In-equality: a critical review of “How Epithets and Stereotypes are Racially Unequal.” Language Sciences, 52, 139-154. Davidson, D. (1978). What Metaphors Mean. Critical Inquiry, 5(1), 31-47. Dipert, R. (1996). Reflections on Iconicity, Representation, and Resemblance: Peirce’s Theory of Signs, Goodman on Resemblance, and Modern Philosophies of Language and Mind. Synthese, 106(3), 373-397. Doherty, W. R. (1997). The Emotional Contagion Scale: Measure of Individual Differences. Journal of Nonverbal Behavior, 21, 131-154. Dror, I. E., Kosslyn, S. M., & Waag, W. (1993). Visual-spatial abilities of pilots. Journal of Applied Psychology, 78, 763-773. Dummett, M. (1981). Frege: Philosophy of Language, 2nd ed. Cambridge, MA: Harvard University Press. Dundes, A. (2002). The Shabbat Elevator and Other Sabbath Subterfuges. Lanham, MD: Rowman & Littlefield. Egan, A. (2008). Pretense for the Complete Idiom. Nous, 42(3), 381-409. Ellinas, A. A. (2013). The Rise of Golden Dawn: The New Face of the Far Right in Greece. South European Society and Politics, 18(4), 543-565. Emanatian, M. (1995). Metaphor and the Expression of Emotion: The Value of Cross-Cultural Perspectives. Metaphor and Symbolic Activity, 10(3), 163-182. Embrick, D. G. & Henricks, K. (2013). Discursive Colorlines at Work: How Epithets and Stereotypes are Racially Unequal. Symbolic Interaction, 36(2), 197-215. Fasoli, F., Maass, A. & Carnaghi, A. (2015). Labelling and discrimination: Do homophobic epithets undermine fair distribution of resources? British Journal of Social Psychology, 54(2), 383-393. Ferber, A. L. (1998). Constructing whiteness: the intersections of race and gender in US white supremacist discourse. Ethnic and Racial Studies, 21(1), 48-63. Frege, G. (1966). On Sense and Reference. In P. Geach & M. Black (Eds.), Translations from the Philosophical Writings of Gottlob Frege (pp. 56-78). Oxford: Blackwell. Frege, G. (1956). The Thought: A Logical Inquiry. Mind, 65(259), 289-311. Frege, G. (1997). Begriffsschrift. In M. Beany (Ed.), The Frege Reader (pp. 47-78). Oxford: Blackwell. Freud, S. (1949). Introductory Lectures on Psycho-Analysis. London: Hogarth. Fridlund, A. J. (1994). Human Facial Expression: An evolutionary view. San Diego: Academic Press. Gadon, O. & Johnson, C. (2009). The Effect of a Derogatory Professional Label: Evaluations of a ‘Shrink’. Journal of Applied Social Psychology, 39, 634-655. Garcia, S. M., Weaver, K., Mozkowitz, G. B., & Darley, J. M. (2002). Crowded minds: The implicit bystander effect. Journal of Personality and Social Psychology, 83(4), 843-853. Geach, P. T. (1965). Assertion. Philosophical Review, 74(4), 449-465. Gendler, T. (2000). The Puzzle of Imaginative Resistance. Journal of Philosophy, 97(2), 55-81. Gendler, T. (2003). On the Relation between Pretense and Belief. In M. Kiernan and D. Lopes (Eds.), Imagination, Philosophy, and the Arts (pp. 125-141). London: Routledge. Gendler, T. (2006). Imaginative Contagion. Metaphilosophy, 37(2), 183-203.

183

Gendler, T. (2008). Alief and Belief. Journal of Philosophy, 105(10), 634-663. Gendler, T. (2011). Imagination. In E. Zalta (Ed.), The Stanford Encyclopedia of Philosophy. Stanford: CSLI publications. http://plato.stanford.edu/entries/imagination/ Gendler, T. and Kovakovich, K. (2005). Genuine Rational Fictional Emotions. In M. Kieran (Ed.), Contemporary Debates in Aesthetics (pp. 241-253). New York: Blackwell. Gibbard, A. (1990). Wise Choices, Apt Feelings: A Theory of Normative Judgment. Cambridge, MA: Harvard University Press. Gibbard, A. (2003). Reasons Thin and Thick. Journal of Philosophy, 100(6), 288-304. Glenberg, A. M., Becker, R. Klotzer, S., Kolanko, L., Muller, S., & Rink, M. (2009). Episodic affordances contribute to language comprehension. Language and Cognition, 1, 113-135. Glezakos, S. (2012). Words Gone Sour? In W. P. Kabasenche, M. O’Rourke, & M. H. Slater (Eds.), Reference and Referring (pp. 385-404). Cambridge, MA: MIT Press. Glickman, N. S. (1996). The Development of Culturally Deaf Identities. In N. S. Glickman and M. A. Harvey (Eds.) Culturally Affirmative Psychotherapy with Deaf Persons (pp. 115- 153). Mahwah, NJ: Lawrence Erlbaum. Goff, P. A., Eberhardt, J. L., Williams, M. J., & Jackson, M. C. (2008). Not Yet Human: Implicit Knowledge, Historical Dehumanization, and Contemporary Consequences. Journal of Personality and Social Psychology, 94, 292-306. Goodman, N. (1968). Languages of Art: An Approach to a Theory of Symbols. Indianapolis: Bobbs- Merrill. Goodman, J., Schell, J., Alexander, M. G., & Eidelman, S. (2008). The Impact of a Derogatory Remark on Prejudice Toward a Gay Male Leader. Journal of Applied Social Psychology, 38, 542-555. Green, L. J. (2002). African American English: A Linguistic Introduction. Cambridge: Cambridge University Press. Green, M. S. (2007). Self-Expression. Oxford: Oxford University Press. Greenberg, J. & Pyszczynski, T. (1985). The effect of an overheard ethnic slur on evaluations of the target: How to spread a social disease. Journal of Experimental Social Psychology, 21(1), 61-72. Grice, H. P. (1989). Studies in the Way of Words. Cambridge, MA: Harvard University Press. Griffin, B. & Smith, H. E. (2013). Autobiography of Mark Twain (Vol. 2). Berkeley: University of California Press. Grushkin, D. (2003). The Dilemma of the Hard of Hearing within the U.S. Deaf Community. In L. Monaghan, C. Schmaling, K. Nakamura, & G. H. Turner (Eds.), Many Ways to Be Deaf: International Variation in Deaf Communities (pp. 114-140). Washington, DC: Gallaudet University Press. Harris, P. L., Brown, E., Marriott, C., Whittall, S. & Harmer, S. (1991). Monsters, ghosts and witches: testing the limits of the fantasy-reality distinction in young children. British Journal of Developmental Psychology, 9, 105-23. Harris, P. L. (2000). The Work of the Imagination. Oxford: Blackwell. Harvey, J. (1999). Civilized Oppression. Lanham, MD: Rowman and Littlefield. Hatfield, E., Cacioppo, J. T., & Rapson, R. L. (1993). Emotional Contagion. Current Directions in Psychological Science, 2, 96-100. Hay, Ryan (2013). Hybrid Expressivism and the Analogy Between Pejoratives and Moral Language. European Journal of Philosophy, 21(3), 450-474.

184

Heyes, C. (2012). Identity Politics. In E. Zalta (Ed.), The Stanford Encyclopedia of Philosophy. Stanford: CSLI publications. http://plato.stanford.edu/archives/win2014/entries/identity- politis/ Hill, T. (1979). Symbolic Protest and Calculated Silence. Philosophy & Public Affairs, 9(1), 83- 102. Holcomb, T. K. (2013). Introduction to American Deaf Culture. Oxford: Oxford University Press. Hom, C. (2008). The Semantics of Racial Epithets. Journal of Philosophy, 105(8), 416-440. Hom, C. (2010). Pejoratives. Philosophy Compass, 5(2), 164-185. Hom, C. (2012). A Puzzle about Pejoratives. Philosophical Studies, 159(3), 383-405. Hom, C. & May, R. (2013). Moral and Semantic Innocence. Analytic Philosophy, 54(3), 292- 313. Hom, C. & May, R. (2014). The Inconsistency of the Identity Thesis. Protosociology, 31, 113- 120. hooks, b. (1995). Killing Rage: Ending Racism. New York: Henry Holt and Co. Hornsby, J. (2001). Meaning and Uselessness: How to Think About Derogatory Words. Midwest Studies in Philosophy, 25, 128-141. Hughes, G. (2006). An Encyclopedia of Swearing: The Social History of Oaths, Profanity, Foul Language, and Ethnic Slurs in the English-Speaking World. Armonk: M.E. Sharpe. Hume, D. (1985). Of the Standard of Taste. In E. F. miller (Ed.), Essays: Moral, Political and Literary (pp. 226-249). Indianapolis: Liberty Fund. Hurston, Z. N. Characteristics of Negro Expression. In C. A. Wall (Ed.) Sweat (pp. 55-71). New Brunswick, NJ: Rutgers University Press. Husemann, L. R. & Taylor, L. D. (2006). The Role of Media Violence in Violent Behavior. Annual Review of Public Health, 27, 393-415. Jahoda, G. & Krewer, B. (1997). History of Cross-Cultural and Cultural Psychology. In J. W. Berry, Y. H. Poortinga, & J. Pandley (Eds.), Handbook of Cross-Cultural Psychology, Theory and Method, Vol. 1 (pp. 1-42). Boston: Allyn & Bacon. Jay, T. (2000). Why We Curse: A Neuro-Psycho-Social Theory of Speech. Philadelphia/Amsterdam: John Benjamins. Jay, T. (2009). Do Offensive Words Harm People? Psychology, Public Policy, and Law, 15(2), 81-101. Jeshion, R. (2013a). Slurs and Stereotypes. Analytic Philosophy, 54(3), 314-329. Jeshion, R. (2013b). Expressivism and the Offensiveness of Slurs. Philosophical Perspectives, 27, 231-259. Johnson, K. (2008). Topic in Ellipsis. Cambridge: Cambridge University Press. Kaplan, D. (2004). The Meaning of Ouch and Oops. Howison lectue in Philosophy. UC Berkeley, August 23rd. Retrived July 11, 2015, from http://eecoppock.info/PragmaticsSoSe2012/Kaplan.pdf Kelly, D. & Roedder, E. (2008). Racial Cognition and the Ethics of Implicit Bias. Philosophy Compass, 3, 522-540. Kendon, A. (2004). Gesture: Visible Action as Utterance. Cambridge: Cambridge University Press. Killmister, S. (2012). Why Group Membership Matters: A Critical Typology. Ethnicities, 12(3), 251-269.

185

Killmister, S. (2014). Resolving the Dilemma of Group Membership. In G. Calder, M. Bessone, & F. Zuolo (Eds.), How Groups Matter (pp. 89-108). New York: Routledge. Kirkland, S., Greenberg, J., & Pyszczynski, T. (1987). Further Evidence of the Deleterious Effects of Overheard Ethnic Labels: Derogation Beyond the Target. Personality and Social Psychology Bulletin, 13(2), 216-227. Kripke, S. (1980). Naming and Necessity. Cambridge, MA: Harvard University Press. Kymlicka, W. (1995), Multicultural Citizenship: A Liberal Theory of Minority Rights. Oxford: Oxford University Press. Lakoff, G. (2012). Explaining Embodied Cognition Results. Topics in Cognitive Science, 4, 773- 785. Lakoff, G. & Johnson, M. (1980). Metaphors We Live By. Chicago: University of Chicago Press. Lampert, M. D. and Ervin-Tripp, S. M. (1998). Exploring paradigms: The study of gender and sense of humor near the end of the 20th century. In W. Rush (Ed.), The Sense of Humor: Exploration of a Personality Characteristic (pp. 231-270). Berlin: Mouton de Gruyter. Lane, H. (2005). Ethnicity, Ethics, and the Deaf-Word. Journal of Deaf Studies and Deaf Education, 10(3), 291-310. Lepore, E. and Stone, M. (2010). Against Metaphorical Meaning. Topoi, 29, 165-180. Leslie, A. (1987). Pretense and representation: The origins of ‘theory of mind’. Psychological Review, 94, 412-426. Lewis, D. (1969). Convention: A Philosophical Study. Cambridge, MA: Harvard University Press. Lewis, D. (1979). Scorekeeping in a Language Game. Journal of Philosophical Logic, 8(1), 339- 359. Liao, S., Strohminger, N. & Sekhar Sripada, C. (2014). Empirically Investigating Imaginative Resistance. British Journal of Aesthetics, 54(3), 339-355. Locke, J. (1995). An Essay Concerning Human Understanding. Oxford: Oxford University Press. Lott, E. (1993). Love and Theft: Blackface Minstrelsy and the American Working Class. Oxford: Oxford University Press. Lycan, W. (1984). Logical Form in Natural Language. Cambridge, MA: MIT Press. Lycan, W. (2015). Slurs and lexical presumption. Language Sciences, 52, 3-11. Mag Uidhir, C. (2013). What’s so bad about blackface? In D. Flory & M. Bloodsworth-Lugo (Eds.), Race, Film, & Philosophy (pp. 51-68). New York: Routledge. Mahar, W. J. (1985). Black English in Early Blackface Minstrelsy: A New Interpretation of the Sources of Minstrel Show Dialect. American Quarterly, 37, 260-285. Marshall, H. & Stock, M. (1958). Ira Aldridge: The Negro Tragedian. London: Rockliff. Matlock, T. (2004). Fictive motion as cognitive simulation. Memory and Cognition, 32, 1389- 1400. May, R. (2005). Bad Words: Remarks on Mark Richard’s ‘Epithets and Attitudes’. USC Workshop in the Linguistics/Philosophy Interface: Syntax and Semantics with Attitude, University of California, Santa Cruz. Maynard Smith, J. & Harper, D. (2004). Animal Signals. Oxford: Oxford University Press. McCready, E. (2010). Varieties of conventional implicature. Semantics and Pragmatics, 3, 1-57. McGinn, C. (2004). Mindsight. Cambridge, MA: Harvard University Press. McIntosh, P. (2000). White Privilege and Male Privilege: A Personal Account of Coming to See Correspondences Through Work in Women’s Studies. In A. Minas (Ed.), Gender Basics:

186

Feminist Perspectives on Women and Men, 2nd ed. (pp. 30-38). Belmont, CA: Wadsworth. McNeill, D. (2005). Gesture and Thought. Chicago: University of Chicago Press. Moran, R. (1989). Seeing and Believing: Metaphor, Image, and Force. Critical Inquiry, 15(1), 87-112. Morris, D., Collett, P., March, P., and O’Shaughnessy, M. (1979). Gestures: their origins and distribution. New York: Stein and Day. Neale, S. (1999). Coloring and Composition. In K. Murasugi & R. Stainton (Eds.), Philosophy and Linguistics (pp. 35-82), Boulder: Westview Press. Nunberg, G. (forthcoming). The Social Life of Slurs. In D. Fogal, D. Harris, & M. Moss (Eds.), New Work on Speech Acts. Oxford: Oxford University Press. Nunberg, G., Sag, I. A., & Wasow, T. (1994). Idioms. Language, 70(3), 491-538. Olsson, C. J., Jonsson, B., Larsson, A., & Nyberg, L. (2008). Motor Representations and Practice Affect Brain Systems Underlying Imagery: An fMRI Study of Internal Imagery in Novices and Active Higher Jumpers. The Open Neuroimagining Journal, 2, 5-13. Ozel, S., Larue, J., & Molinaro, C. (2004). Relation Between Sport and Spatial Imagery: Comparison of Three Groups of Participants. The Journal of Psychology: Interdisciplinary and Applied, 138(1), 49-64. Padden, C. & Humphries, T. (1990). Deaf in America: Voices from a Culture. Cambridge, MA: Harvard University Press Padden, C. and Markowicz, H. (1997). Learning to be Deaf: Conflicts Between Hearing and Deaf Cultures. In M. Cole, Y. Engstrom, & O. A. Vasquez (Eds.) Mind, Culture, and Activity: Seminal Papers from the Laboratory of Comparative Human Cognition (pp. 418-431). Cambridge: Cambridge University Press. Payne, B. K. (2006). Wepon Bias: Split-Second Decisions and Unintended Stereotyping. Current Directions in Psychological Science, 15(6), 287-291. Pettigrew, T. F. & Tropp, L. R. (2006). A meta-analytic test of intergroup contact theory. Journal of Personality and Social Psychology, 90(5), 751-783. Potts, C. (2007). The Expressive Dimension. Theoretical Linguistics, 33(2), 165-198. Putnam, H. (1975). Mind, Language and Reality: Philosophical Papers Volume 2. Cambridge: Cambridge University Press. Queen, R. (2007). Sociolinguistic Horizons: Language and Sexuality. Language and Linguistic Compass, 1(4), 314-330. Reeve, C. D. C. (1997). Plato, Cratylus: translated with introduction and notes. Indianapolis: Hackett. Richard, M. (1993). Articulated Terms. Philosophical Perspectives, 7, 207-230. Richard, M. (2008). When Truth Gives Out. Cambridge, MA: Harvard University Press. Ritchie, K. (2015). The Metaphysics of Social Groups. Philosophy Compass, 10(5), 310-321. Rozin, P. Millman, L. & Nemeroff, C. (1986). Operation of the Laws of Sympathetic Magic in Disgust and Other Domains. Journal of Personality and Social Psychology, 50(4), 703- 712. Saka, P. (2007). How to Think About Meaning. Berlin: Springer. Schlenker, P. (2007). Expressive Presuppositions. Theoretical Linguistics, 33(2), 237-245. Schwitzgebel, E. (2002). How well do we know our own conscious experience? The case of visual imagery. Journal of Consciousness Studies, 9, 35-53.

187

Sennet, A. & Copp, D. (2015). What Kind of Mistake is it to Use a Slur? Philosophical Studies, 172(4), 1079-1104. Shachar, H. (2014). The Holocaust is not your metaphor to use in modern political debates. Retrived March 13, 2016, from http://www.theguardian.com/commentisfree/2014/jan/27/the-holocaust-is-not-your- metaphor-to-use-in-modern-political-debates Shakabpa, T. W. D. (2010). One Hundred Thousand Moons: An Advanced Political History of Tibet. Translated by D. F. Maher. Brill: Leiden. Smiley, M. (2010). Collective Responsibility. In E. Zalta (Ed.), The Stanford Encyclopedia of Philosophy. http://plato.stanford.edu/entries/collective-responsibility/ Soames, S. (1989). Presupposition. In D. Gabbay & F. Guenthner (Eds.), Handbook of Philosophical Logic, Volume IV: Topic in the Philosophy of Language (pp. 553-616), Kluwer Academic Publishers. Sperber, D. & Wilson, D. (1986). Relevance Theory: Communication and Cognition. Oxford: Blackwell. Stalnaker, R. (1999). Context and Content: Essays on Intentionality in Speech and Thought. Oxford: Oxford University Press. Stenico, C. & Greitemeyer, T. (2014). The Others Will Help: The Presence of Multiple Video Game Characters Reduces Helping After the Game is Over. Journal of Social Psychology, 154, 101-104. Stenner, A. J. (1981). A Note on Logical Truth and Non-Sexist Semantics. In M. Vetterling- Braggin (Ed.), Sexist Language: A Modern Philosophical Analysis (pp. 299-306), New York: Littlefield, Adams and Co. Stern, B. (1988). Figurative Language in Services Advertising: The Nature and Uses of Imagery. Advances in Consumer Research, 15, 185-190. Strawson, P. F. (1950). On Referring. Mind, 59(235), 320-344. Thoutenhoofd, E. D. (2000). Philosophy’s Real-World Consequences for Deaf People: Thoughts on Iconicity, Sign Language and Being Deaf. Human Studies, 23, 261-279. Toll, R. C. (1974). Blacking Up: The Minstrel Show in Nineteenth Century America. Oxford: Oxford University Press. Tomasello, M. (2008). Origins of Human Communication. Cambridge, MA: MIT Press. Tuomela, R. (1989). Actions By Collectives. Philosophical Perspectives, 3, 471-496. Turner, R. N. & Crisp, R. J. (2010). Imagining intergroup contact reduces implicit prejudice. British Journal of Social Psychology, 49, 129-142. Van Leeuwen, N. (2013). The Meanings of ‘Imagine’ Part 1: Constructive Imagination. Philosophy Compass, 8, 220-230. Waldron, J. (2012). The Harm in Hate Speech. Cambridge, MA: Harvard University Press. Walton, K. (1978). Fearing Fictions. Journal of Philosophy, 75, 5-27. Walton, K. (1990). Mimesis as Make-Believe. Cambridge, MA: Harvard University Press. Walton, K. (1994). Morals in Fiction and Fictional Morality (I). Proceedings of the Aristotelian Society, 68, 27-50. Wharton, T. (2009). Pragmatics and Non-Verbal Communication. Cambridge: Cambridge University Press. Whiting, D. (2007). Inferentialism, Representationalism and Derogatory Words. International Journal of Philosophical Studies, 15(2), 191-205.

188

Whiting, D. (2013). It’s Not What You Said, It’s the Way You Said It: Slurs and Conventional Implicature. Analytic Philosophy, 54(3), 364-377. Williams, B. (1985). Ethics and the Limits of Philosophy. Cambridge, MA: Harvard University Press. Williamson, T. (2009). Reference, Inference and the Semantics of Pejoratives. In J. Almog & P. Leonardi (Eds.), The Philosophy of David Kaplan (pp. 137-158), Oxford: Oxford University Press. Williamson, T. (2010). The Use of Pejoratives. In D. Whiting (Ed.), The Later Wittgenstein on Language (pp. 45-62). Oxford: Oxford University Press. Wilson-Mendenhall, C. D., Simmons, W. K., Martin, A. & Barsalou, L. W. (2013). Contextual processing of abstract concepts reveals neural representations of nonlinguistic semantic content. Journal of Cognitive Neuroscience, 25(6), 920-935. Windley, M. (2013). Adam Goodes ‘gutted’ after 13-year-old girl’s racial slur, who called the Sydney champion today to apologize. Retrived March 13, 2016, from http://www.heraldsun.com.au/sport/afl/adam-goodes-gutted-after-13-year-old-girls- racial-slur-who-called-the-sydney-champion-today-to-apologize/story-fni5fan7- 1226650256245 Young, I. M. (1990). Justice and the Politics of Difference. Princeton, NJ: Princeton University Press. Zack, N. (2002). Philosophy of Science and Race. New York: Routledge. Zwaan, R. A., Stanfield, R. A., & Yaxley, R. H. (2002). Language Comprehenders Mentally Represent the Shapes of Objects. Psychological Science, 13, 168-171.

189