Plurk paste

PART III
Idealization

CHAPTER SEVEN Neutrality We must take sides. Neutrality helps the oppressor, never the victim.
—ELIE WIESEL, NOBEL PRIZE ACCEPTANCE SPEECH1 TIME AND AGAIN, in this book, we have argued that central concepts in the theory of meaning, for example presupposition, are characterized in a way that is not adequate to the generality of the phenomena they were introduced to explain. And a theme running through these pages has been that this characteristic failure of generality is a consequence of problematic idealizations in the study of speech. In this part of our book, we raise this underlying theoretical vein of our book to the surface.
In this chapter, and the next, we single out two idealizations for extended critique. These are the idealizations of neutrality and straight-talk.
We have chosen these two idealizations for several reasons. First, because the objections to them are different in character, developing both gives us a broader sense of the ways in which idealizations can distort. Secondly, we have chosen these concepts for their broader importance to democratic political philosophy. The problems we raise for them arise from the study of speech. However, the problems these idealizations contribute to range considerably wider. Finally, in chapter 9, the last chapter in this methodological section of the book, we place critiques of idealizations in the theory of meaning into the larger context of debates about “ideal theory” in other areas of philosophy.
This chapter concerns the ideal of neutrality. Our conclusion is a kind of deliberate misreading of the quote from Elie Wiesel at the start of this chapter. The quote suggests and justifies a moral imperative, echoing in stark terms a long history of warnings of the dangers of neutrality. We instead interpret “We must take sides” metaphysically: it is impossible to be neutral. More narrowly, the main thesis we defend is that words can never be neutral.
We began this book with the concepts of resonance, attunement, and harmony. These three concepts can be seen, respectively, as generalizations of meaning, belief, and logical consistency. Words have various resonances.
The resonances allow communication of attunements, and harmonization settles the attunements of individuals and groups into coherent wholes. The resonances may be rich, including all the elements of ideologies. A language is a web of resonances that gives flesh to its speakers’ social identities. We will argue that the model of speech we have built upon these conceptual generalizations is in tension with natural language ever being neutral.
Among other things, we deny that speech can ever lack all but denotational resonances. If you idealize away from resonances that are nonneutral (in the sense we discuss), you idealize away from important linguistic reality. In the case of neutrality, we argue that the idealizations are especially distorting, misrepresenting underlying metaphysics.2 7.1. A Neutral Space for Reasons?
Let us say, provisionally, that discussion is neutral if perspective and attunement to social location are irrelevant to the understanding and evaluation of each move in the discussion.
Here is a natural way of thinking of how propaganda impairs democracy. Democracy is conceived as a space of neutral deliberative reason. In such a space, participants in a discussion solely focus on exchanging reasons. Since the space of deliberation is neutral, it is devoid of biased perspectives. Since it is devoid of biased perspectives, the best argument wins. Why, according to this picture, do we not live in a democracy? Because propaganda prevents the neutral exchange of reasons.
Propaganda wields perspective as a weapon, sometimes explicitly, other times covertly, masking perspective behind a facade of apparent neutrality.
Propaganda excites emotion and fosters in-group bonding, impeding rationality. Propaganda is thus a threat to the realization of the democratic ideal. Propaganda and ideological discourse are barriers to the neutral space of reasons that is the liberal ideal.3 Such a position is naive, because of course even the most committed public-reason liberal can grant that there are important appeals to emotion, for example, to address gaps in democratic culture. In a more sophisticated vein, one could allow that some propaganda—some appeals to emotion, to a persuasive force not backed by reasons—could be necessary to achieve democratic ends.
4 When some of the population is in the grips of ideology, reasons will not work to free them.
Ideology can block considerations of reason.
On this picture, even for democratic ends, propaganda is sometimes necessary to shock people out of ideological barriers. But it is only as a way station to the democratic ideal of neutral deliberative reason. Speech that resonates emotionally with people, bonding them via emotions such as empathy, can perhaps play a role as a means to realizing democratic ideals.
But ultimately, the idea goes, a true democratic space of reasons does not involve speech whose function is to resonate with an audience via means other than the provision of reasons. Public reason is, on this picture, incompatible with perspectival content.
Such a neutral conception of public reason is not required even for highchurch epistemic democracy. W. E. B. Du Bois’s essay “Of the Ruling of Men” is an epistemic argument for democracy. Du Bois argues that only by including everyone’s voice will one devise good policy: Continually, some classes are tacitly or expressly excluded [from democratic participation]. Thus women have been excluded from modern democracy because of the persistent theory of female subjection and because it was argued that their husbands or other male folks would look to their interests. Now, manifestly, most husbands, fathers, and brothers will, so far as they know how or as they realize women’s needs, look after them. But remember the foundations of the argument—that in the last analysis only the sufferer knows his sufferings and that no state can be strong which excludes from its expressed wisdom the knowledge of its mothers, wives, and daughters. We have but to view the unsatisfactory relations of the sexes the world over and the problem of children to realize how desperately we need this excluded wisdom.5 Du Bois’s argument is explicitly epistemic, but it concerns highly subjective perspectives. Du Bois’s argument is that we need the insight from highly subjective perspectives for good policy, and no amount of learning or good intentions will suffice to make policy without them.
Though the idea of a space of neutral “unsullied” public reason is not necessary to defend some kind of version of deliberative democracy, it nevertheless proves tempting as a model.
In commenting on the arguments of Catharine MacKinnon and Rae Langton to regulate pornography, Judith Butler asks, “If pornography performs a deformation of speech, what is presumed to be the proper form of speech? What is the notion of nonpornographic speech which conditions this critique of pornography?”6 Butler adds dryly, “Though neither Langton nor MacKinnon consults Habermas, their projects seem to be structured by similar cultural desires”; Butler is suggesting that their projects require an idealized speech community in which the only consideration is (to use Habermas’s phrase) “the force of the better argument.”7 The temptation to appeal to an ideal of neutral public reason is easy to slip into, even when one does so unawares.
In Alice Crary’s 2018 paper, “The Methodological Is Political: What’s the Matter with ‘Analytic Feminism’?” Crary critiques this picture of the antidemocratic nature of emotional or antirational appeals in public discourse. In particular, she criticizes it for presupposing a picture of idealized public discourse as a neutral space of reasons. Using as an example the account of propaganda in Stanley’s How Propaganda Works, since it is for her a paradigm case of how not to proceed, Crary questions whether it is reasonable for feminists to view radically antiestablishment methods that undercut enlightenment-style rationalism as a short-term tactic en route to a new and better “neutral conception of reason.” She views such an approach as deeply flawed, and says of its advocates: They recognise the practical need for ethically non-neutral methods.
At the same time, they claim that these methods are as such nonrational and should therefore only be used—as crucial but also intrinsically problematic and therefore merely temporary instruments —for clearing away obstacles to the creation of a space for debate that is maximally neutral and, as the thinkers in question see it, hence rationally and politically sound. This is the stance that Jason Stanley, for instance, defends in his recent, widely discussed book on propaganda. Despite regarding all propaganda as non-neutral and hence as non-rational and politically problematic, Stanley allows for indispensable or, in his terms, “non-demagogic” types of propaganda that are at times “necessary” for dismantling ideological formations that distort what he sees as the neutral space for democratic discourse.8 In the framework developed in this volume, the notion of “a neutral space of reasons” for discourse is incoherent and epistemologically problematic.
It is incoherent, because utterances of words unavoidably are tied to speech practices that locate speakers in histories and social roles. Perspective is inherent to communication.
If our model of speech is correct, the ideal of a neutral space of reasons is also epistemologically problematic. It is epistemologically problematic because, given that ought implies can, taking it as an ideal suggests it is possible to avoid thoroughly unavoidable communicative effects of speech practices. It is, therefore, an ideal that occludes reality. It is a dangerous ideal.
Our model undermines the ideal of neutral reasoning. But perhaps this just shows that our model of speech is problematic, insofar as it cannot make sense of the notion of a neutral space of reasons. In what follows, we will critique the motivations and indeed the very coherence of this ideal. In the end, it emerges as a virtue of our model that it cannot make sense out of it, rather than a problem.
What would ground the ideal of a “neutral space for democratic discourse”? One would need to think of linguistic meaning as having a certain kind of neutrality, that language has a nonideological core that is the subject of semantic and pragmatic theory. According to this containmentmetaphor-based model, the “core” of meaning is some sort of neutral, objective information, with a denotation independent of the speaker.
But first—what is “neutral, objective information”? Our discussion here is necessarily speculative, as we are not committed to the coherence of the notion. We begin with its source in the analytic tradition—the work of Gottlob Frege. Frege’s preferred notion of meaning, which he called sense, was a kind of neutral informational content. Frege recognized that there were aspects of meaning that were not neutral in his preferred sense.
But he rejected these as irrelevant to his “official” notion of meaning. We will argue that these aspects are conventional, and we can understand them as such. 7.2. Frege on Sense versus Tone In Frege’s mature philosophy of language, from 1892 on, there are two levels of meaning: sense and reference. The reference of a proper name is an object, so “Mark Twain” and “Samuel Clemens” have the same reference. But, famously, Frege argues that in such cases the two names may have different meanings, as it is possible to believe that Mark Twain is not Samuel Clemens, without irrationality (or, to return to the classic example that Frege used, that Hesperus is not Phosphorous). These different meanings, the different meanings associated with “Mark Twain” and “Samuel Clemens,” Frege calls senses. Here are five doctrines Frege holds about senses: 1. The sense of an expression is the way the expression presents its referent.9 “The Evening star” and “The Morning star” present the same referent, Venus, in different ways.
2. The sense of an expression is what one must grasp in order to understand the expression.10 3. The sense of an expression is its reference in opaque, or indirect, contexts, such as in the ‘…’ of “John believes …” 4. Sense determines reference, at least in the minimal sense that a difference in reference entails a difference in sense.
5. Some senses, the thoughts, are the ultimate objects of truth and falsity.
On this picture of meaning, each word has a meaning, the way it presents its referent, and the thought expressed by the sentence is composed of the meanings (senses) of the words the sentence contains. The thought is what is believed, and it is what is judged true or false.
In “Logic” and elsewhere, Frege draws a distinction between sense and what he calls tone.
From the distinctions he draws between sense and tone, we can draw a conclusion about Frege’s views of senses. It is that senses are neutral and aperspectival: If we compare the sentences “This dog howled the whole night” and “This cur howled the whole night,” we find that the thought is the same. The first sentence tells us neither more nor less than does the second. But whilst the word “dog” is neutral as between having pleasant or unpleasant associations, the word “cur” certainly has unpleasant rather than pleasant associations and puts us rather in mind of a dog with a somewhat unkempt appearance.11 Frege’s argument that the two sentences tell us the same thing is that they have the same truth conditions, namely, they contain the same information about the world. Consider someone who objects to the view that the two sentences “This dog howled the whole night” and “This cur howled the whole night” give us the same information, holding that the second sentence also expresses the information that the speaker holds a negative view toward the dog. Frege writes, We assume that the first sentence is true and the second sentence is spoken by someone who does not actually feel the contempt which the word “cur” seems to imply. If the objection were correct, the second sentence would now contain two thoughts, one of which was false; so it would assert something false as a whole, whilst the first sentence would be true. We shall hardly go along with this; rather the use of the word “cur” does not prevent us from holding that the second sentence is true.12 Frege holds that senses of words are the elements relevant to the truth or falsity of what is said by sentences containing them. Since tone is not relevant to truth or falsity, it is not part of sense. This Fregean picture has developed into the view that the core of the theory of meaning is the study of meaning properties relevant to truth or falsity of the assertion. Since the differences between “cur” and “dog” are not relevant to truth and falsity of assertions, they are not relevant to the theory of meaning.
From our perspective, Frege’s argument is problematic because it privileges the semantic predicates “true” and “false” over other predicates of communicative acts, and by focusing on these predicates, it also implicitly restricts the purview of the theory of meaning to acts of assertion and the truth conditions of sentences that are asserted. Speech does not merely attune people cognitively, to properties of the external world, but also attunes them emotionally, socially, and dispositionally. Speech that attunes people to emotions can be wounding or inspiring. When it also attunes people to the world, it can simultaneously be true or false. Given that speech can attune us to emotions and practices, we should not from the outset privilege the metasemantic predicates “true” and “false.” There is another thought articulated here by Frege that is natural and widespread. It is that the “normal” words, like “dog,” are “neutral as between having pleasant or unpleasant associations.” The idea that a theory of meaning can ignore tone receives a kind of illicit support from the fiction that nonneutral words have neutral counterparts. Do they?
7.3. Evaluative Predicates and Value-Laden Concepts Many concepts are inherently evaluative. Consider concepts from virtue ethics, like the concepts of generosity or cruelty.
These “thick” moral concepts mix descriptive and evaluative elements, and even perhaps affective ones. The practice of calling an action “generous” is a practice of praising that action. The practice of calling an action “cruel” is a practice of condemning it. The reason there is no neutral counterpart of “generous” is not merely that the term involves perspective, but that it involves positive valuation. It is not clear how any development of Frege’s notion of tone could be of much use here. It seems to us at the very least unhelpful to say that the sense of “generosity” is neutral, but that it has a negative tone.
Few would want to suggest that words like “cruel,” “vindictive,” “selfish,” and even words for “thin” moral concepts like “right” and “wrong,” have neutral counterparts. But this is surely not just a matter of language, not just a question of there happening to be certain words that are used evaluatively. Surely, if people can be said to have concepts corresponding to words like “cruel” and “generous,” then they have concepts that are intrinsically evaluative. It’s not merely that our words are neither neutral nor built on a neutral core, but that our ways of categorizing the world are neither neutral nor built on a neutral core.
We do not attempt to give in this volume a theory of concepts or categories. But we will say this. We can see no reason to assume that the way people categorize is in any sense neutral, and there also is no reason to assume that categorization is in some sense restricted to purely epistemic attunements. Whether the notion of a purely epistemic attunement is coherent is not obvious to us, but we would certainly want to allow that a person’s categories might involve perceptual, emotional, and dispositional attunements.
People’s concepts of everything from cotton wool and knives to Bach fugues and jokes may be inseparable from how they perceive them, how they feel about them, and how they use them. But if at least some concepts that words pick out are not in any interesting sense neutral concepts, but are rather intrinsically laden with value, then it is not clear what it would mean for a word that has such a concept as its meaning to have a neutral counterpart. The neutral counterpart would have to be a concept that is not value-laden. But then what would make that concept the neutral counterpart of the original?
We have no idea what you get if you neutralize the concept of being selfish, but we think it would of necessity lack the things that make the concept of being selfish what it is; it would not be a neutral version of the original concept, but rather a different concept altogether. This suggests to us that for at least some words, not only do they lack neutral counterparts, but they could not possibly have neutral counterparts. In such cases, to restrict attention to a neutral core would be to lose almost all of the fruit.
There is, then, good reason to say that some words lack neutral counterparts. Whatever the relationship between “mom” and “mother,” it is not that the latter corresponds to the neutral core of the former. Neither the word “mother” nor the concept of motherhood is neutral, either ideologically or emotionally. But let us go further and consider a case in which the existence of a neutral counterpart has been taken as obvious. If the neutral counterpart hypothesis can be argued to fail even here, then the case for neutral counterparts as a useful component of semantic theorizing is on rocky ground indeed.
7.4. “Dog” versus “Cur” The Fregean example of cur is suggestive, but misleadingly so. “Cur” is not merely rare, but essentially never used in contemporary English with the sense of dog + <negative affect>, except by philosophers of language. If it is so hard to come up with an example of a word conveying tone that the standard examples are themselves vanishingly rare, then how important can tone be? Contrary to the impression that the example “cur” gives, we suggest that tone is ubiquitous.
Let us start with the supposedly neutral counterpart “dog,” which we choose precisely because of its apparent innocuousness. An unpretentious word, certainly. Old English. And clearly the least marked noun to describe a dog, or at least the most obvious word to use.
But neutral? What would it even mean for it to be “neutral”?13 Consider first the fact that the etymological history of English words correlates to the register in which they are used. Thus, those Old English words, like “child” and “chip” and “chin,” which are still in common use, tend to be understood as ordinary colloquial, plain English words. AngloNorman vocabulary is more varied in its register, with “beef,” and “blue,” and (in the United Kingdom) “bucket” being broadly distributed across registers, but “citizen,” “commodity,” and “conspiracy” carrying a decided edge of specialization, education, and privilege. And more obviously Latinate or Greek vocabulary presumably provides many of the five-dollar words in Mark Twain’s aphorism “Don’t use a five-dollar word when a fifty-cent word will do.” Such language tends to sound cultured or learned —from “amorousness” and “abdominal” to “phobias” and “philosophy.” The point here is not that these etymologies provide any hard and fast rules as to the extrasemantic significance of words. Rather, etymology provides a way of sorting words into heaps, and once the words are in those heaps, broad differences in usage and connotation become obvious. Simply categorizing words by length, or by frequency, would achieve almost the same as sorting by etymology.
Let us now propose that “dog” is inherently tonal, introduce natural objections, and then at least partially rebut them. It is obvious first that “dog” is often used in phrases that have negative connotations. Although there are certainly expressions in which “dog” is used positively, the negative resonances in English are mirrored by the fact that the “recurring themes in common idioms in languages such as French, German, Italian, and Spanish, are those of: low status/worthlessness, futility, unhappiness, competition/aggression.”14 Such associations surely creep into many uses of “dog.” And “dog” being, as noted, an unpretentious Old English word, will also carry its unpretentiousness with it wherever it goes.
One might, of course, use or mention the word “dog” in a pretentious utterance, but that is true of even the most mundane function words like “in” and “a”: one should not expect an unpretentious word to significantly reduce the pretention of a pretentious sentence any more than one should expect lightweight clothing to lower the weight of an overweight wearer.
A number of objections surface at this point. One might accept that when one uses the word “dog,” various connotations connected with the attitudes and social positioning of the speaker are available, but deny that those have to do with the meaning of the word. One might, perhaps, say that some of the connotations reflect nothing about the meaning of “dog” but rather its usage. And one might say that yet other connotations have nothing to do with the word “dog,” but rather are associations we have with the concept represented by “dog,” associations we could in principle have even if we didn’t know the word “dog.” And one might note that “dog” is often used without any conscious intention to display unpretentiousness, or, for that matter, a negative attitude toward “dog.” Doesn’t this suggest that if there is any such coloration, it is not part of the conventional meaning of the word?
Let us tackle the last of these first. Does the potential lack of conscious intention imply a lack of convention? Although, as we will see, the possibility of unconscious invocation of tonal meaning is significant, we do not take that possibility to directly bear on the issue of whether or not a given tonal coloring is a conventional aspect of a word’s meaning. At risk of an argument from authority, let us borrow a line of thought with such philosophical pedigree that it is part of the canon of undergraduate philosophy of language: Putnam’s externalism, as previously invoked in section 5.2.15 A use of “beech” conveys the meaning of “beech” rather than “elm” even if the speaker does not have a sufficiently advanced conceptualization of either beeches or elms to distinguish between the two, for example, being unaware of the fact that the base of beech leaves is more symmetric than that of elm leaves. If the speaker lacks conceptual distinctions, then clearly the communicative intention cannot involve conscious access to those distinctions. Therefore, the argument would go, one can utter something with a certain meaning without necessarily having conscious access to all aspects of that meaning. And a fortiori, one can utter something that conventionally carries a certain tonality without conscious awareness of that tonality. Just as one can, as the schoolyard rhyme suggests, be a poet and not know it, so can the tone of one’s words reveal piety or pride, and cause offense or delight, all despite any lack of conscious intention. Even the words of a tone-deaf speaker may carry tone.
The framework we developed in parts I and II is, among other things, one way of developing Putnam’s approach to externalism. Where he justified an externalist account of meaning in part on the basis of the existence of experts who could in principle pass judgment on the proper application of botanical terms like “beech” and “elm,” we instead depend on collective attunement in a community of practice. We suggested, in chapter 2, that collective attunement of a group to a practice does not imply full individual attunement by the members of the group. So, for us, what differentiates the resonances of “beech” from the resonances of “elm” is the differences in the way they are both used within a community of practice, that community being, roughly, the English-speaking world, or any large and highly connected subpart of it.
Furthermore, we defined the resonances of an action in such a way that they do not essentially depend on the intentions of a particular actor, but on the extension of the practice that the action instantiates, an extension that is defined relative to a community of practice.
On our approach, it would be possible for some practice to develop that crucially depended on intention, and indeed there can be practices for which the primary communicative function is to indicate one’s purpose, but, equally, a practice can become conventionalized in such a way that intentions are not central or strong resonances. To clarify: what we mean by saying that a resonance is not central is that there are other stronger resonances, and what we mean by saying that a feature of context is not a strong resonance of a practice is that an action instantiating that practice is only associated with a low probability boost for that feature. Given our probabilistic approach, even when a certain type of intention is associated with a practice, it will not follow that the intention is present at every or even most performances of the practice. To return to an example that has been an intuition pump throughout this book, one can analyze the functioning of the practice of saying “ouch” without supposing that an intention to express pain is a strong resonance, although such an intention may be analyzed as a weak resonance, present on some occasions of use but not others. What of the idea that what we have identified as the potential tonal components of “dog,” the unpretentiousness and a somewhat demeaning attitude toward dog-like things, are not part of the meaning of “dog”? It is intuitive to analyze the claimed negativity as being an attitude toward doglike things, rather than part of the meaning of the word. And it seems plausible to say that the unpretentiousness of “dog” is no more and no less than a fact about usage, associated with our knowledge of register rather than with our knowledge of the meaning of the word. We have two replies, one somewhat defensive, and the other accepting.
Our more defensive reply is that we cannot see a clear empirical basis for discriminating tone as part of expression meaning from tone as either associated with an underlying nonlinguistic conceptual category, or from tone as a matter of usage. We can, on the contrary, see at least one argument for considering at least some of the tonality of “dog” a part of its meaning.
We can think of this using Grice’s notion of detachability. Detachability is a diagnostic criterion for whether an aspect of meaning is conventional or conversational, although, as we will discuss later, it is by no means clear that the criterion is a coherent one. Grice describes an inference associated with an expression as “nondetachable insofar as it is not possible to find another way of saying the same thing (or approximately the same thing) which simply lacks the [inference].”16 So let us ask: is the unpretentiousness of “dog” detachable?
The word “canine” used as a noun (or indeed the Latin Canis familiaris) would presumably be taken by Grice as having the same extension as “dog,” but contrast strongly in pretentiousness. So it is possible to find another way of saying “the same thing (or approximately the same thing)” that lacks the inference in question, namely the inference that the speaker is acting in an ordinary, humble, or unpretentious way. It is less clear whether the words “canine” or “hound” carry similar negative connotations as “dog,” though certainly they are not found in the same negative idioms. We do not live in a hound eat hound world. It anyway seems that, by Grice’s criterion, at least the inference of unpretentiousness is detachable from the concept represented by “dog,” and therefore this inference should be considered to be conventionally associated with the word “dog.” Here is a fact: use of one and the same word on different occasions can exemplify different speech practices. This simple fact has ramifications for a test like detachability. There are regular speech practices, involving use of “dog,” that have negative connotations. Use of the phrase “dog eat dog world” is a clear case. There are other speech practices, which uses of “dog” also can exemplify, that lack any obvious negative connotations, and that may indeed have positive connotations. The occurrence of the word in the phrase “dog show” is a clear case.
The word “dog” is associated with all of these speech practices. Which speech practice a particular use of “dog” exemplifies is communicatively relevant. We have here many conventional associations, rather than none.
Communicative actions are relevant for coordination in a particular context. The word “dog” is typically used as part of a practice of unpretentious speech; it fits into a way of speaking that is appropriate in particular contexts. Whether a particular use of “dog” exemplifies a practice of this kind is a context-dependent matter. The context-dependence of practices is built into our model, much like the context dependence of indexicals. We have not exploited much of the power of probability theory in our account so far, restricting ourselves to the correlative connection between an action and a feature. But the same mathematical framework allows arbitrarily more complex dependencies. In principle, the resonances of a single practice could involve a high probability of feature A whenever feature B is present in the context, and a low probability of feature A whenever feature B is absent. The question for us would not be whether a meaning can in principle be like that, but whether such a meaning is likely to become conventionalized and remain somewhat stable in a community of practice. That is, there is an empirical question of the extent to which members of a speech community categorize a range of actions involving a complex dependency of this sort as belonging to a single practice, and the extent to which, contrarily, the speech community will categorize the range of actions as belonging to two distinct practices, one used in contexts where B is present, and one where B is absent.
A word exemplifies a speech practice or practices, the practices that are salient in the context in the sense that the presupposed resonances of the practice are a good match for that context. Being a good match implies both that in such contexts there would be a significant probability of the practice being performed, and that the practice does not have strong positive resonances with features absent from the context, or strong negative resonances with features that are present. And it is the practices that are salient in the context, those it is recognizable as exemplifying, that determine the communicative effect of uttering it (including its referential content, which, as Kripke emphasized, is determined by speech practice).
The examples we have discussed lead us away from a model of communication that privileges a notion of neutrally descriptive denotation as central to meaning. If the speaker and hearer both know that Mary has exactly one pet and that it’s a Labrador, then in saying “Mary’s X is missing her,” many values of X would lead the speaker and hearer to become jointly attuned to the same property of the world, for example, X = “dog,” “doggywoggy,” “puppy-wuppy,” “damned dog,” “hound,” “pet,” “Labrador,” or “lab.” That two expressions attune conversational participants to the same properties of the world is consistent with these expressions having all sorts of other communicative effects. What other sorts of linguistic phenomena suggest a similar moral?
7.5. Speech Practices In The Language of the Third Reich, Klemperer describes the propaganda of the Third Reich under which he lived. The speech practices that were pervasive under National Socialism, the subject matter of the book, were sufficiently distinctive as to give them a name, Lingua Tertii Imperii, or LTI: The LTI only serves the cause of invocation.… The sole purpose of the LTI is to strip everyone of their individuality, to paralyze them as personalities, to make them into unthinking and docile cattle in a herd driven and hounded in a particular direction, to turn them into atoms in a huge rolling block of stone.17 Klemperer’s book is a focused description of LTI, of characteristic Nazi speech practices.
The first chapter of Klemperer’s book, “Heroism: Instead of an Introduction,” is devoted to describing the symbols associated with German cognates of the term “heroism,” what he describes as the “uniform,” in fact the “three different uniforms,” of the word. The first uniform was that of the “blood-soaked conquerors of a mighty enemy,” the image of the original stormtroopers of the 1920s. The second uniform was that of “the masked figure of the racing driver,” representing German success at the beloved sport of auto racing. The third uniform was that of the wartime tank driver.
These are the symbols that the term “heroism” evoked. In all three cases, the symbols were “closely tied up with the exaltation of the Teutons as a chosen race: all heroism was the sole prerogative of the Teutonic race.”18 Klemperer writes, What a huge number of concepts and feelings [the Nazi cast of mind, the typical Nazi way of thinking] has corrupted and poisoned!
… I have observed again and again how the young people in all innocence, and despite a sincere effort to fill the gaps and eliminate the errors in their neglected education, cling to the Nazi thought processes. They don’t realize they are doing it; the remnants of linguistic usage from the preceding epoch confuse and seduce them.
We spoke about the meaning of culture, or humanitarianism, of democracy and I had the impression that they were beginning to see the light, and that certain things were being straightened out in their willing minds—and then, it was always just round the corner, someone spoke of some heroic behavior or other, or of some heroic resistance or simply of heroism per se. As soon as this concept was even touched upon, everything became blurred, and we were adrift once again in the fog of Nazism … it was impossible to have a proper grasp of the true nature of humanitarianism, culture, and democracy if one endorses this kind of conception, or to be more precise misconception, of heroism.19 Klemperer is here saying that there is a distinctive National Socialist speech practice. In this speech practice, only Aryans are described as “heroic.” Use of the word “heroic” attunes an audience raised under National Socialism to a practice of treating Aryans as better than non-Aryans. Indeed, the function of the speech practice of heroism myths in Nazi ideology is to attune audiences to such a practice. Use of the term “heroic” also attunes the audience to emotions, resonating positively with images of Storm Troopers and race-car drivers, and (presumably) negatively to images of cosmopolitan decadence, to homosexuality, and to swarthy Slavic or Semitic faces. Klemperer claims that those raised under National Socialism react to the use of the term “heroic,” and similar vocabulary, inadvertently —the use of the term leads the audience to have positive emotional attunements to certain images and negative emotional attunements to others. A Nazi propaganda poster from 1940, set in a characteristic gothic font and with emphatic capitalization, attributed the following quote to Hitler: “HEROISMUS ist nicht nur auf dem Schlachtfelde notwendig, sondern auch auf dem Boden der Heimat” (Heroism isn’t only needed on the battlefield, but also on the soil of the homeland). In Nazi ideology, “heroism” was part of an Aryan supremacist speech practice. It became a conventionalized property of “heroism,” or at least of the German “Heroismus,” to be a constituent of such a speech practice.
If we think of using words as exemplifying speech practices, we can understand how they can have conventional significance over and above the contents they are used to present. A use of the word “heroism” in Germany in the early 1940s conveyed a wealth of associations, in virtue of the Nazi speech practices that such a use exemplified. A noun like “stormtrooper” was used to predicate a property of some individual. A certain set of individuals were in the extension of the predicate, namely the stormtroopers. One of the things the term “stormtrooper” could be used to do was predicate membership of this set.
Philosophers sometimes talk of the “ordinary denotation” of an expression, although this is not part of our own technical development. For such philosophers, the ordinary denotation of “stormtrooper” would be the property of being a stormtrooper, which they may or may not have distinguished from the property of being in the extension of the term “stormtrooper.” For some, following the line most clearly developed by Rudolf Carnap and Richard Montague, one aspect of the denotation, its intension, would be a somewhat more sophisticated variant of this property.
The notion of intension plays a similar role in contemporary theory of meaning as Frege’s notion of sense played in his account, although Frege himself didn’t provide any very clear analysis of what senses are, beyond that they are modes of presentation. On the Carnap/Montague view, the intension would not just be a set of individuals, but a property that varies depending on how the world is, essentially a function from the way the world is to the set of individuals that would be in the extension of “stormtrooper” if the world was that way. No such notion of denotation can adequately capture what is expressed by “stormtrooper.” What “stormtrooper” expresses is not just membership of a set, but placement within a complex social world seen from a certain perspective. And use of “stormtrooper” was not just expressive. The use of “stormtrooper” in Nazi Germany helped make the Nazi world what it was, in virtue of the practices its use exemplified.
There are many mechanisms that play a role in forming an ideology.
With Jack Balkin in his theory of “cultural software,” we focus on symbolic forms, which, in his words, “carry units of cultural transmission.”20 Nazi ideology embodies a way of thinking of Aryans and a way of thinking about Jews, and these ways are carried in the speech practices that constitute it.
Nazi ideology is an extreme example of an oppressive ideology. It brings out in vivid ways how ideologies can unjustly create hierarchies of worth. But any way of talking about other beings brings with it attunement to emotions, images, and suggested practices. If one is asked to “go out with the guys,” one can make a free choice about one’s decision—but the language in which the invitation is offered will elicit emotions that will influence one’s choice.
The emotions will be different if one asked to “meet with mutual acquaintances,” even if the shared references are the same.
Ignoring the force of such emotions, the result of the framing of an invitation, is missing a vital function of language. It is a function of language that is missed if one only focuses on neutral shared references of words.
Let us pause to recall (from section 2.6) that in our descriptive sense of “ideology,” not all ideologies are oppressive. An ideology helps to guide us through a complex world by giving us shortcuts and strategies in the face of an overwhelming mass of information. An oppressive ideology involves shortcuts that impose or reinforce hierarchies of ethnicity, gender, or other dimensions that should not be set into hierarchies of value. A liberatory ideology involves practices that undercut or disrupt such hierarchies. And some ideologies are just habits, practices, and concepts that help an agent maneuver the world in mundane, practical ways—to find food, for example.
As we have seen in the previous section, we do not need to move to extreme cases, such as National Socialist speech practices, to see that words can conventionally signify much more than the extensional properties of the world they are used to present. This is a fact about conventional significance that is far more general. When speakers use words like “dog,” they also exemplify speech practices that are located within ideologies and manifest a particular perspective. The words “stormtrooper” and “dog,” although nouns of very different stripes, are alike in respect of both having the trivial property that their resonances are not neutral.
We can call the property trivial for a simple reason: not only are resonances never neutral, but it is entirely unclear how it could possibly be the case that a practice could arise that had neutral resonances. Practices arise within communities of practice. At a minimum, a practice carries as a resonance the ideological perspective and social location of the practitioners.
There is no such thing as a neutral practice. The idea of a neutral practice makes no more sense than the idea of a neutral ideology.
The function of an ideology is to create a set of presupposed shared meanings—not just common beliefs, but attunements to more than just the beliefs that we associate with the entities, events, actions, and properties under discussion. A full theory of language should allow us to model these presuppositions. That is, a theory of language should allow us to model not just the commonly shared beliefs that our words presuppose, but commonly shared attunement to emotions and practices associated with ways of referring to objects, actions, and events. It should allow us to model the degree of attunement that reflects differential attachments to these ideologies. Oppressive ideologies clearly frame the social world with charged emotions—but even casual conversation persuades along dimensions that cannot simply be captured by considering “neutral” shared references.
7.6. Perspective In sketching the neutral picture of content within standard theory, we began with the work of Gottlob Frege and his distinction between sense (neutral content) and tone. The picture Frege sketches in 1897 neglects indexicality, words like “I,” “here,” “now,” “this,” and “that,” whose reference shifts with context. A central moral of the last half century of discussion of indexicality is that it seriously compromises any putative commitment to neutrality of meaning. Any defense of neutral meaning against our skepticism must reckon with this moral, which is, as it were, internal to the ideology of analytic philosophy of language and mind. John Perry famously argued in “Frege on Demonstratives” that the moral of indexicality is that nothing could play all of Frege’s five roles of senses.21 Frege seems committed to the hypothesis that the linguistic meaning of the first-person pronoun “I” is always the same. Given that sense is supposed to be linguistic meaning, this means that “I” has the same sense in every context. For Frege, a sentence expresses a thought by virtue of how the senses of the words in the sentence combine. So, under the hypothesis that the sense of “I” is constant, if President Obama were to utter in 2015 “I am the President of the United States,” he would express the same thought as the one David Beaver would express by using that sentence at the same time. But then the thoughts would have the same truth value.
However, the thought President Obama would have expressed would be true, and the thought David Beaver would have expressed would be false.
Therefore, the thoughts are different after all. Frege must abandon one or more of the roles of sense.
In the end, Frege comes around to recognizing the failure of the picture of sense he had earlier addressed. When he addresses the topic of indexicals in his paper “The Thought,” he decides that the word “I” has different senses in different contexts (his discussion shows clear commitment to role 4, that a difference in reference entails a difference in sense).22 He abandons the publicity of sense, on the grounds that each person’s first-personal way of thinking of themselves is accessible only from that person’s perspective.
The sense of an occurrence of “I” is only accessible from one perspective. It is not a sense that is aperspectival, and so, as John Perry and others have long pointed out, Frege’s own discussion violates the thesis of the neutrality of meaning. On Frege’s preferred view, our first-person thoughts cannot be grasped by others.
Frege’s initial discussion of first-person thought, the de se, has had a tremendous impact on the theory of meaning. It could rightly be said that the problem of perspective is one of the perennial issues of the last century in the theory of meaning in the analytic tradition. There is a lengthy tradition of defending Frege’s view that first-personal thoughts are not shareable.
As Gareth Evans argued, Frege’s view that first-person perspective is not sharable is consistent with Frege’s commitment to the objectivity of thoughts—in the sense that a first-person perspective is an objectively existing perspective, albeit only capable of being grasped by one person.23 Evans embeds his defense of Frege’s position in Frege’s theory of meaning, employing senses rather than modal semantics. But Frege’s view of irreducibly unshareable first-person thoughts is not bound to this theory of meaning. David Lewis took the notion of an irreducibly first-person perspective so seriously that he altered his basic conception of content, moving from propositions as sets of possible worlds to propositions as sets of centered worlds, worlds centered upon a first-person perspective.24 Lewis’s technology is baked into the notion of resonance we have developed, since resonances involve probabilities of features occurring in a certain type of centered context. Building on Lewis’s work, Robert Stalnaker has argued for the coherence and importance of “a notion of informational content that is not detachable from the situation of a subject, or from a context in which the content is ascribed.”25 In short, the thought that informational content is irreducibly perspectival is deeply embedded inside the analytic tradition in the theory of meaning.
Suppose that what is meant by perspective is not a generalized notion like that introduced in the last chapter (the attunements to a set of features of context that some individual or group has), but is rather something like Lewisian content. Let us term such a notion “indexical perspective.” Then one may, in an effort to achieve something like neutrality, attempt to detach perspective from discourse by appeal to a language stripped of indexicality, say for the purposes of science. Addressing the historian’s attempt to conceal perspective from their language, Roland Barthes writes, As a matter of fact, in this case, the speaker annuls his emotive person, but substitutes it for the “objective” person: the subject subsists in its plenitude, but as an objective subject; this is what Fustel de Coulanges called, significantly (and rather naively), “the chastity of history.” On the level of discourse objectivity—or lack of signs of “the speaker”—thus appears as a special form of image repertoire, the product of what we might call the referential illusion, since here the historian claims to let the referent speak for itself. This illusion is not proper to historical discourse—how many novelists— in the realistic period—imagine they are “objective” because they suppress signs of “I” in the discourse! … We know the absence of signs has a meaning too.26 Discourse that appears to lack perspective is often a way of masking a perspective.27 Even if one doubts that indexical perspective is ubiquitous in empirical discourse, it is nevertheless very often present when it appears not to be.
It has been widely recognized that irreducibly perspectival content poses challenges to idealizations about rationality and communication, how to retain and build on a core of information over time, and how to share that information with others. In short, the modifications in the theory of meaning required to incorporate the ubiquity of irreducible and shareable perspectives pose challenges to the notion of a neutral deliberative space of reasons for shared rational inquiry. There are many creative solutions—for example, Sarah Moss has argued that to each perspectival proposition (de se proposition), there corresponds a nonperspectival (de dicto) proposition that can “rationally stand in” for that perspectival content in communication and learning.28 Moss’s solution of finding a de dicto “stand in” for each de se proposition (like the approach of Stalnaker29) relies on denying a core motivation intuition of the de se literature—that two people (Lewis’s “two Gods”) can share all their de dicto beliefs, while differing on their de se attitudes. It seems to be generally conceded, by such solutions, that differences in individual perspectival contents must supervene on differences in ordinary (de dicto) contents. If individual perspectival content runs as deep as Lewis and Evans hold, then it does raise significant problems for how to make sense of neutral debate. Andy Egan’s work on the problem of the de se is an attempt to alter the framework of the theory of meaning in more substantial ways than Moss, not by finding a de dicto (“neutral”) counterpart of de se belief, but by relativizing the fundamental notions of the theory of meaning, such as the common ground of a conversation and (accordingly) presuppositions, both of which can have de se content. Different roles involving the same de se common ground (speaker, or hearer) result in grasping different propositions (the speaker grasps a proposition about themselves when uttering, “I am tired”; the addressee grasps a proposition about the speaker, while both speaker and hearer have de se attitudes toward their joint common ground, relativized to themselves).30 This is a distinct way to alter the theory of meaning to account for something like shared discourse information—but the information that is shared, on this account, is not the same. Only components of that information are shared, as when you and we are thinking of different scenes involving the same number of daffodils.
The moral of the extensive literature on individual perspective over the past two decades is that we are already stuck with a difficult problem of objectivity at the very heart of orthodoxy in the theory of meaning. Just the problem of the de se alone raises significant conceptual difficulties for the concept of a neutral space of reasons for deliberation.
Andy Egan’s theoretical interest in the de se was a small part of his contribution to the large literature on relativism, a major topic of research in philosophy of language in the last two decades. Perhaps the signal work of this era is John MacFarlane’s Assessment Sensitivity.
31 An assessmentrelative theory of meaning is one that takes norms governing notions like agreement or disagreement to be relative to a context (or standard) of assessment. If there is such a thing as a neutral debate, it would presumably have to involve agreement or disagreement about the same propositions.
The central theoretical task facing any assessment-relative account is to make sense out of such agreement and disagreement, in other words, to make sense out of the fundamental concepts of rational debate.
Most work on assessment sensitivity is driven by examples in which a robust sense of objectivity is in any case misplaced. For example, one of the central motivating examples for this revision to the theory of meaning is predicates of personal taste, as in “is tasty.”32 These are domains in which it is easier to see how one might abandon commitment to a genuine neutral space of reasoned debate. However, we might imagine, as in the case of the de se, that the problem of social perspective is more general, and indeed ubiquitous.
Oddly, given the centrality of the problem of the de se, and the recent exploration of the conceptual foundations of assessment sensitivity, analytic philosophers working in the theory of meaning have been more reluctant to explore the consequences of theorizing about social perspective as a node or element in the theory of meaning. The goal of revising the theory of meaning to accommodate an irreducibly first-person perspective has been central to almost every tradition in the analytic theory of meaning. But there is considerably less discussion of how to alter the theory of meaning to account for social perspective, and less discussion of the sorts of linguistic phenomena that would raise the issue in the first place. As we have seen, there does not seem to be a legitimate objection in principle to such a program, as subsequent work on assessment-relativism has shown.
The literature on assessment sensitivity concerns how truth is evaluated.
But the notion of a social perspective need not be theorized as a parameter of truth assessment. Elisabeth Camp has shown that perspective does not need to be theorized simultaneously with the evaluation of truth, or as a question of how a predicate “is true” is applied to sentences or propositions.
In a pioneering set of papers, she has argued that perspectives are implicated in “political discourse, intimate interpersonal arguments, informal commentaries on movies—anywhere that intuitive interpretation is at stake.”33 Camp’s perspectives are comparable to the landscapes of attunement introduced in the analysis of harmonization in section 3.4, determining what concepts and dispositions tend to be activated together in a person’s thought, and they can be seen as a direct antecedent of our proposal.
Camp has applied her account to a range of issues. She suggests that perspectives are central to understanding essential features of metaphors, in particular their lack of paraphrasability.
34 She has also argued that perspectives are necessary to understanding how slurs derogate—by introducing derogating perspectives.35 (We will return to the treatment of slurs in chapter 10.) In short, over the years, Camp has given a systemic case, considering a variety of different linguistic constructions and phenomena, for the importance of perspective as an essential tool in the theory of meaning.36 Suppose someone wished to maintain the possibility of democratic debate as involving a neutral space of reasons. They would not only need to counter worries arising from arguments that point of view is inherent in the very notion of a propositional claim, but they would also need to sidestep the worry that every claim is associated with some perspective, some way of thinking about the claim involving implicit priorities about what to think about. It is entirely unclear what it would even mean for a perspective, in Camp’s sense, to be neutral. But let us not here make the contentious claim that every utterance involves a perspective in Camp’s sense, or a social perspective of any sort. For the purposes of this section, what we take from Camp’s work is the more minimal claim that she provides a tangible demonstration that it is coherent to develop a theory of meaning within which social perspectives might play a central role. Social perspective is often conveyed in narrative form, or with respect to background narrative frames, as discussed in section 3.5.37 Full interpretation of speech requires a background of narratives, which must augment the “simple picture” that assigns objects, properties, and relations to expressions. It is against a background of narrative that we interpret speech, not vice versa. But if this is so, then what would neutral speech be?
Could it be speech in which the underlying narratives were neutral? If the concept of a neutral story makes any sense at all, we doubt that a neutral story would be any good. But independently of this value judgment, neutral narrative does not seem to us a fruitful direction on which to ground a more general notion of neutral speech, for the concept of a neutral narrative is at best ill-defined, and at worst incoherent. It is unclear what it would mean for a story to be neutral, to be a way to frame patterns of events that is free of ideology.
Could neutral speech could be speech that is entirely bereft of narratives? Let us suppose for a moment that speech bereft of narrative is a coherent concept. Let us also leave aside what seems to us a natural view, that speech bereft of narratives would be far from ideal, and would in fact be tragically impoverished. The sense in which speech bereft of narratives would not be neutral is simple. Far from being neutral, narrativeless speech would be a particular and rather narrow subspecies of discourse. It would be exceptional and extraordinary. It would be speech that did not support narrative modes of thought like those we discussed in chapter 3, modes of thought that many have taken to be central to human culture (cf. our discussion in section 3.5 of narrative harmonization).38 Preferring narrativeless speech to more common modes of human discourse would require a particular ideological position. We should call it an extreme position. Whether extreme or not, it would not in any sense be a neutral position. Hence the hypothesized narrativeless speech would also not be neutral. It would be the speech of ideologues who fancied their speech better than that of the bulk of humankind.
Narratives are not the only structures against which we interpret strings of sentences; we can add to this list stereotypes, value systems, perspectives, prejudice, framings, practices, identity, and affect. Each of these notions is unlikely to be straightforwardly reducible to another. Jessie Munton has recently argued that prejudice is not adequately captured by false beliefs about a group, negative affect, or prejudicial behavior.
39 According to Munton, prejudice toward a group is rather a problematic way of organizing information about that group—one that gives certain features more weight than they should receive. Prejudice, on Munton’s view, is a problematic salience structure, one that makes certain features of group members more salient than they ought to be. If what I say can presuppose prejudice, and prejudice is a salience structure, then perhaps, to model speech, we need to think of presupposition as a relation, not just to information, but to something psychologically richer.
We have presented an account of presupposition that makes this concrete. In the model of presupposition developed in chapter 5, a practice can presuppose any system of attunements that is commonly present among the participants in the practice. Further, as we argued in section 3.4, attunements are organized into landscapes that encode activational relationships. Thus one attunement may selectively activate others, and hence determine a default flow of attention, as when, absent external stimuli to redirect attention, one follows through steps of an argument, a story, or a plan. Our model allows such structures to be presupposed by a practice, provided that the presence of the structure is in a position to affect whether the practice is performed (the causal efficacy postulate of section 2.5). We believe the model thus allows us to capture the idea present in prior work (in particular, in the work of Camp and Munton) that in presupposing a stereotype, we do not merely presuppose a set of facts, but also presuppose tendencies to focus on some characteristics rather than others.
We could, perhaps, attempt to individually add each of the major notions we have mentioned in this section—narratives, framings, prejudice, identity, etc.—to an ultimate theory of communication, characterizing each as we go. However, our hope is that by this point in the book the charitable reader is starting to get a feel for how such developments might be advanced. Perhaps the reader is even sympathetic to the possibility that the model we have proposed already does some of the work that might be demanded of these notions. Rather than picking through each of the many further notions that we have discussed, let us try to consider a broader context in which they can be seen as playing their own roles. The broader context is that of ideology, which we defined in chapter 2 as follows: Ideology: An ideology is the system of collective attunements among members of a community of practice.
The term “ideology” is used in a bewildering variety of senses, as befitting the fact that the study of ideology has been central to philosophical theorizing dating back at least to Plato.
Raymond Geuss defines the broadest sense of ideology, “ideology in the descriptive sense,” as including “the beliefs the members of a group hold, the concepts they use, the attitudes and psychological dispositions they exhibit, their motives, values, predilections, works of art, religious rituals, gestures, etc.”40 In this sense, ideology is not pejorative—every group has an ideology—but rather is something like a worldview, understood broadly enough to include practices. As emphasized already in this chapter, our definition of ideology is ideology in such a descriptive sense.
Pierre Bourdieu writes, “Every established order tends to produce (to very different degrees and with very different means) the naturalness of its own arbitrariness.”41 This is a central function of an ideology. An ideology creates a set of presuppositions—what Bourdieu calls “doxa,” “the class of what is taken for granted … the sum total of the theses tacitly posited on the hither side of all inquiry.”42 Among the doxa in Bourdieu’s sense—the presuppositions imposed by ideology—we include more specific notions, such as narratives, stereotypes, practices, and so forth. This too is how Bourdieu thought about these notions—as Judith Butler writes, emphasizing the materiality of doxa, realized as practices and habits, “ ‘Rules of the game’ are, quite literally, incorporated, made into a second nature, constituted as a prevailing doxa.”43 Across very different paradigms in the theory of meaning, philosophers have held that first-person perspective is something that must be modeled, to account for an essentially dominant view that first-person perspective is ubiquitous and theoretically central.44 We have seen a two-decade attempt to make sense more generally of disagreement about less than factual seeming domains. There is thus no obstacle in principle that theorists of meaning in the analytic tradition have had to the centrality and importance of perspective—indeed, they have been theorizing it as a parameter of truth.
We have argued in this section that once it is already granted that the theory of meaning must be adjusted to account for first-person perspectives, there is no obstacle to exploring the ways in which perspectives, narratives, and, overall, ideologies, structure communication.
7.7. Against Neutrality as an Ideal The thought behind neutrality is that the core of communicated denotation is nonideological. On this way of understanding neutrality, the core information conveyed in a communicative action is neutral because it does not in any way convey something about social perspective.
But, as we have argued, ideologies are constituted in part by social practices. And speech practices are paradigmatic examples of social practices. In The Language of the Third Reich, Klemperer writes, Nazism permeated the flesh and blood of the people through single words, idioms and sentence structures which were imposed on them in a million repetitions and taken on board mechanically and unconsciously.… Language does not simply write and think for me, it also increasingly dictates my feelings and governs my entire spiritual being the more unquestioningly and unconsciously I abandon myself to it.45 Klemperer here forces us to take seriously the thought that National Socialist ideology consists in large part of certain speech practices. It is for this reason that denazification took the form of constraining and minimizing those speech practices.46 The result of denazification was a more open, inclusive, and democratic German society. Denazification is an example of restrictions on speech practices leading to a strengthening of democratic norms.
Social practices include “cultural schemas,” such as “rules of etiquette, or aesthetic norms, or such recipes for group action as the royal progress, grain riot, or democratic vote, or a set of equivalences between wet and dry, female and male, nature and culture, private and public.”47 A speech practice embodies a schema in this sense, a way of speaking of the world, characteristically the social world, in a way that assigns different groups social roles and positions. In so doing, a speech practice reflects a perspective on the world. These speech practices thus help constitute ideologies—social perspectives on the world (according to one reading of Elisabeth Camp’s work, speech practices are the result of such perspectives, which give rise to them).
If we think of a use of a word as occurring within a speech practice, then we can see that there will standardly be communicative effects of using that word over and above attuning an audience to a common object, property, or relation. By using one word rather than another for the same thing, “dog” versus “canine,” for example, we locate ourselves within one speech practice rather than another. The speech practices our words exemplify locate us socially, and often politically. Different speech practices can use some of the same tools—that is, can employ some of the same words. One and the same word can exemplify, in one context one speech practice, and in another context, quite a distinct one. So, different uses of one and the same word can communicate quite different things, depending on the speech practice they exemplify. But this is not the lack of linguistic convention—it is participation in multiple linguistic conventions.
The connection between presupposition and practice makes particularly vivid the problems with a model of speech that focuses on neutral debate as an ideal. Words are always embedded in speech practices, which are elements of ideologies. All speech will locate its users in this way. If we ignore this fact, or pretend it is something we can idealize away from, we are liable to overlook the effects of the speech practices that words exemplify. The use of a word can cue us to the social location of a speaker, to their identity, their socioeconomic class, gender, or race. We might say that the ideal of neutrality is problematic insofar as it masks the way that speech essentially presents the world within a situated perspective.
Though we take this to be true, it does not quite go far enough, for it focuses on the limiting nature of the ideal of neutrality for assertions, that is, communicative acts whose primary purpose is to present the way the world is. Before considering other communicative acts, let us turn briefly to one of the most important lines of epistemological work in recent years.
As we have noted, not all ideologies are oppressive. Some standpoints bring with them epistemological advantages, which is the central thesis of standpoint epistemology. There are various versions of standpoint epistemology. There are some according to which the thesis is a truism— being in a certain social situation leads you to have more fine-grained knowledge about it, for example. There are some according to which it is a weighty metaphysical claim. For example, if certain social situations can be transformative, in the sense of Laurie Paul, they can give you access to information that you cannot have without that experience.48 This is a weighty standpoint epistemological claim. If standpoint epistemology is true, then certain social perspectives can aid you evidentially. Since speech practices are parts of social perspectives, ways of talking can presuppose ways of thinking about the world that are more epistemologically advantageous.
All communication takes place with respect to a context of practices, which licenses the communicative acts constitutive of the communicative exchange. Understanding the interplay between utterances and different background ideologies is essential to understanding what is communicated and to which audience. On our model, there is no sense to be made, as long as one is using words in a living language, of the possibility of reasoning free of a set of background practices. And such situated background practices affect the rhetorical force of one’s chain of reasoning. If standpoint epistemology is correct, knowing that someone is speaking from a privileged epistemic perspective can even be relevant to deciding whose views carry the most weight. From this more realistic perspective, the very coherence of an ideal of neutrality is questionable. Let us strengthen that. Considerations from standpoint epistemology suggest that the coherence of an ideal of neutrality is questionable as applied to knowledge, and equally to the act of assertion. What of other communicative acts? As we have discussed in this volume, borrowing heavily from the tradition of work that runs through Wittgenstein and Austin, people do much in communication besides asserting things about the world. People greet each other, express feelings of pain, happiness, or love, sentence others to jail, joke, threaten, promise, pray, offend, and thank. When at a Black Lives Matter march people in unison repeat “Sandra Bland! Say her name!” they are not asserting anything, and neither is anything being asserted when at a Donald Trump rally people repeat “Lock her up!” The ubiquity of nonassertive speech acts in ordinary discourse, as well as explicitly political discourse, leads to a significant problem for the ideal of neutrality. And here it does not even matter whether we talk of “ideal” in the sense of idealization, that is, theoretical abstraction or simplification, or “ideal” in the sense of preferred way of doing things. The problem is not with the word “ideal,” but with the word “neutral.” The problem is that communication consists of actions drawn from practices. Just as it is unclear what a neutral perspective would be, it is also entirely unclear whether there is any coherent notion of a neutral action or of a neutral practice. Take the mundane practice of greeting. What one might be tempted to describe as a neutral greeting would, in many contexts, appear standoffish for its lack of effusiveness. No greeting is neutral. One cannot joke neutrally, pray neutrally, or sentence someone to death neutrally any more than one can cross a road or kill a fish neutrally. It is simply a category error to apply the term “neutral” to the act of crossing the road or killing a fish. One can at best cross the road or kill a fish without manifesting one’s feelings, but that is not to act neutrally in any interesting sense, and it is not a sense of neutrality that would seem to be of great interest either to philosophers of language or political theorists.
There is another sense in which an action can be neutral, as when someone acts as a “neutral arbiter.” In this sense, to call the action “neutral” is to say that the actor gave no net benefit to any of the parties in the affair being arbitrated. This concept of neutrality is prominent in political debate.
It is related to what is meant by talk of a country’s neutrality in a war, although in this case something weaker might be meant: that a government’s public actions do not so decisively favor the victory of one side that these actions should themselves be considered acts of war.
It is the also sense most relevant to the Wiesel quote that “Neutrality helps the oppressor” with which we headed this chapter, which concerns people’s moral obligation to speak up when systematic oppression is growing around them. Silence may be golden, but it is the oppressor who has a need to trade in its currency, for an absence of speech is an absence of dissent.
We see then that just as speech is never neutral, so it is with silence.
Silence registers tacit acceptance of the staus quo. But here let us note that in The Origins of Totalitarianism, while Hannah Arendt does see silence as an important aspect of the onset of totalitarianism, she develops a more nuanced view than Wiesel, grounded in a broader historical study of totalitarian movements in the twentieth century. The problem is not only silence in the face of overt acts of oppression such as beatings on the street and firing of Jewish workers in 1930s Germany, but the more general failure of people to speak up when national government systematically overlooks, or even supports, gross inequality. In describing the circumstances within which totalitarian regimes grow, she observes that “democratic government had rested as much on the silent approbation and tolerance of the indifferent and inarticulate sections of the people as on the articulate and visible institutions and organizations of the country.”49 That is, the silent acceptance of oppressive systems and the forced silence of the oppressed are, for Arendt, properties of an unjust society, whether superficially democratic or not. Such a society is ripe for totalitarian takeover. Under the guise of giving the oppressed and powerless a voice, totalitarians create a chaos of noisy and violent dissent in democracy in order to displace democracy in favor of an oppressive system that better suits them.
To describe a speech act as neutral in the sense of not benefiting specific sides in an adversarial situation is to presuppose some very specific distinctions, such as who the relevant parties are, and what constitutes harm or benefit. That is, to arbitrate neutrally in this sense is to act in a way that a particular ideology determines is neutral. But if neutrality of a speech act is to be relativized to an ideology, then it is of limited practical or theoretical interest. Practically, an ideology-dependent measure of neutrality cannot serve as a very general tool for arbitration between ideologies.
Theoretically, to accept neutrality as a relative notion is to accept the impossibility of the project of defining a coherent notion of neutrality with the qualities that it was purported to possess, allowing words in some sense to stand above the fray, elevated from the whims of mortals, impartial as the stars above. In some religions, the word of sacred texts is understood to have such qualities, but for anyone standing in a tradition of postenlightenment rationalism, the ideal of neutrality cannot possibly be so lofty.
There is, as we have seen, a substantial literature on the topic of rescuing agreement and disagreement in the presence of various kinds of perspectival relativity. One solution, taken in the assessment-relativity literature, is to limit the scope of claims of perspectival relativity to cases in which there isn’t a sense of shared objective reality—a paradigm example being judgments of taste.50 The terrain looks different in the case of social perspective—one cannot focus on a limited set of contexts in which social perspective is irrelevant (as we have urged). One might think this raises a profound problem for this investigation. One might, that is, think that the ubiquity of social perspective means that we must compromise on the existence of robust truth claims.51 Fortunately, this is not so. Though social perspective is ubiquitous, it need not be a parameter of truth assessment, as seen in Camp and Munton’s work.
In our account, the possibility of a robust theory of truth rests on the fact that attunement is not limited to social categories and practices: there is also attunement to physical properties of the world. Within the community of practice that makes the exchange meaningful in the first place, uptake of a true claim about the world will result in attunement to the world as it actually is, and not merely as it happens to be understood.
There is an analogy to be drawn here with perception, and with Gibsonian direct theories of perception in particular.
52 Modes of perception differ dramatically, with sight, touch, and hearing revealing different qualities of the world in different ways. Furthermore, one individual’s hearing and neurological processing may differ arbitrarily from another’s.
Perception is not neutral, whether one is wearing rose-tinted spectacles or not. Yet different individuals with their different perceptual apparatuses may become attuned to the same physical facts in the world, say the proximal onset of a storm. Similarly, and although communicative practices vary arbitrarily, the statement “It’s going to rain” can be true when uttered in a particular context by virtue of the fact that uptake of the claim by a member of the English-speaking community following the practices of that community would attune that person to a property of the world. Every claim is ideologically relative, and claims about the weather are no exception, but that does not preclude there being a fact of the matter as to whether a claim made by a member of a specific community of practice on a particular occasion is true or not.
To be clear about our position, we are not claiming that for any particular class of statements it is practical, or even possible, to robustly assess the truth value. The existence of statements for which it is possible to robustly assess truth clearly does not imply that the truth of any significant class of statements can be robustly assessed in this way. What we have in mind here are not just issues of vagueness and paradox that have dogged philosophy for millennia, as might be seen in the question of whether it is raining when a single drop falls. We are thinking also of cases in which aspects of ideology, and their constituent practices, are contested. For example, any claim involving a gendered term, including commonplace terms like “man” and “woman,” enters upon ideologically contested territory.
There are then at least three problems to overcome in assessing the truth of a claim involving contested terms. First, there is the ancient problem of vagueness. In our terms, this means that for some claims it may be impossible to determine sufficient history of practice for all the constructions used in the claim such that it is clear whether the physical nature of the world and the conventions of the community of practice determine the claim to be true, or determine it to be false. Second, there is the problem of contestation, which implies that different people may see the same claim through different ideological lenses, with the implication that even though there may be a fact of the matter as to whether a particular statement is true within a particular community of practice, people may contest the truth of a sentence because they do not share collective attunement to the same ideology, and may even sometimes intentionally contest the truth of a sentence as a proxy for contesting ideology. Third, there is a problem of irreducibility of perspective. By this we mean that even if there are particular properties of the world that would determine whether a certain statement within a particular community of practice were true, there could not possibly be a neutral way of paraphrasing the truth conditions of the statement.
Regarding the third issue, the point is that paraphrasing involves reframing, but not deframing. Paraphrasing produces a new symbolic object that one must interpret with a specific ideology. One may well have clarified in the process, perhaps shifting a complex issue into more easily understandable sub-issues, but in doing so one has not removed perspective.
In fact, one has potentially muddied the waters, since the process of paraphrasing introduces ideology in two ways, first thorough the intrinsically ideological practice of paraphrasing itself, and second through the ideological practices that make the constructions in the paraphrase meaningful. This is so whether the paraphrasing is into natural language (the paradigmatic case being the famous instance of Tarski’s truth schema “ ‘Snow is white’ is true iff snow is white”53), or into the mathematical terms of contemporary formal semantic theory.
The practice of labeling certain claims as true, we submit, is connected to the propensity of those claims to attune people to properties of their environment, both properties of the world and properties of their speech community. At the same time, the practice of labeling things as true may also (much like the practice of labeling things as neutral or objective) play other roles, like the role of pushing or enforcing the priority of an epistemic standpoint. Developing a resonance-based theory of the practice of labeling things as true is not a project we take on in this volume. Quite generally, we think that while plenty of attention has been paid to the nature of truth, insufficient attention has been paid to the role of nonpropositional aspects of language.
While we have emphasized that there is much at stake in communication beyond truth, our proposal is not at all to do away with the notion of truth in favor of, say, a radically relativistic view on which the truth of claims is just a personal or even community preference. The relativization we are urging is not a relativization of the sort discussed in the assessment-sensitivity literature, where truth can be taken relative to a parameter of taste or a potential future history. We hope to help explain attributions of predicates like “arrogant,” or “sad,” to an assertion, attributions that reflect socioemotional attunements people may have to the assertion—not to explain how attributions of truth are more complex than they appear to be.
We have argued that speech attunes us to ideologies, which harmonize narratives, practices, affect, values, principles, goals, and expectations, so that there is no “simple” attunement to things. Further, we have deemphasized the role of truth in the theory of meaning: nowhere in this book do we offer a definition of a truth predicate in terms of our model of meaning. The fact that we do not focus on the nature of truth only reflects our opinion that it has been overemphasized in much prior work; it does not reflect any conviction that defining truth would be a mistake. However, both our denial of simple attunement to things and our de-emphasis of truth may lead some to a legitimate worry. The worry would be that the path taken in this book, and in this chapter in particular, is toward a relativist abyss in which all communicative acts are epistemologically equal, none cleaving more tightly to reality than any other. We do not believe that our commitment to speech as pervasively ideological forces us to sacrifice a kind of robust realism. Attunement to ideology does not preclude attunement to reality. On the contrary, attunement to sophisticated ideology, for example scientific theory, can support far more nuanced attunement to reality than would otherwise be possible. While the question of what reality consists in is complex, since culture and cultural objects are part of it, we hold that it is always a legitimate question whether a particular communicative act better attunes people to reality or not. Furthermore, we doubt that as complex a communicative system as a human language could possibly have evolved if it did not have at least a strong tendency to help attune interlocutors to reality. It is not the case that realism is in tension with the ubiquity of social perspective, and realism about communication does not depend on there being such a thing as neutral speech.
To make this discussion more concrete, consider an example. Suppose that just as you are starting to cross the street, an onlooker who has observed an oncoming truck shouts, “Watch out!” and you leap back to the safety of the sidewalk. There is no interesting sense in which the onlooker’s panicked exhortation was neutral, and we would not normally describe such a speech act as true or false. Fortunately, it did better attune you to reality.
So it is with all communication: whatever social, emotional, and dispositional attunements a particular communicative act brings, the question of whether the act is suited to the context in which it is used is a practical one. Among the ways in which an act can be suited to its context is the way it attunes interlocutors to physical aspects of that context. We submit that it is quite possible for a theory of communication that eschews neutrality to nonetheless support a robust realism, because the theory will distinguish the contexts in which acts attune people to reality from those in which they do not.
The considerations in this chapter suggest that there are significant problems with the concept of neutrality. Let us consider what follows from that. First, are particular claims about specific stretches of discourse being neutral coherent, or are they false? Second, to the extent that the notion of neutrality is coherent enough for such claims of neutrality to be false, could it nonetheless be the case that the ideal of neutrality is one worth aiming for, even if it is never actually attained? The answer to both of these questions depends on what notion of neutrality is being assumed. If what is meant is discourse that lacks any ideological perspective—what might be termed “ideological neutrality”—then any claim of neutrality would be incoherent, and indeed the ideal of neutrality would be incoherent too. Discourse occurs in a language, and a language is embedded in an ideology, so discourse cannot be free of ideology. Furthermore, a claim of ideological neutrality consists only in a failure to recognize that alternative perspectives could add anything to the discussion. So such a claim is not only incoherent, but itself cannot possibly be part of a neutral discussion. To claim ideological neutrality is to claim epistemic and communicative supremacy over groups that do not adhere to the speaker’s ideology. Such claims are thus examples of what are often termed “power moves,” paradigmatic examples of a communicative action for which social location is relevant.
Let us return to the provisional definition of neutrality of a discussion from section 7.1, whereby neutrality demands that “perspective and attunement to social location are irrelevant to the understanding and evaluation of each move in the discussion.” Let us allow that perspective need not be completely irrelevant, because we might reasonably say that a primary goal of discussion is to change someone’s perspective. That is, according to the definition of perspective given in section 6.4, the goal of discussion might be to change someone’s attunements to a set of features of the context.
Let us instead limit what must be irrelevant to presupposed perspective: while an effect of a neutral discussion might be to change perspective, perspectival presuppositions and social location should be irrelevant. But even with this limitation of perspective to what is presupposed, perspective will always be relevant to understanding the moves in a discussion, for the simple reason that there can be no understanding of discussion about features of context without prior perspective on at least some of those features. Furthermore, as we have argued, all language has presupposed resonances of social location, starting with membership of the interlocutors in common communities of practice, without which there could not be mutual understanding of what is said.
We can go further along this line: if we are going to idealize about discussion, then, to put it in our own terms, we might say that if an ideal discussion were going to achieve anything, it would be collective harmonization. Harmonization is an increase of collective attunement, such as collective attunement to the world (or “common knowledge,” to use the standard term), and this brings with it the possibility of enhanced future coordination. If one of the functions of an ideal discussion is to create collective attunement, and thence enhanced coordination, then it is hard to understand how social location could ever be irrelevant.
While we have not considered every possible definition of neutrality, we have considered a range of issues related to the concept of neutrality in this chapter, and they draw us to the following conclusions. First, if neutrality is to be coherent at all, it must be a limited notion.
Ideological neutrality is incoherent, and neutrality might at most be applied to a subset of moves in discussions, those that describe the world, and not to speech acts in general.
Second, when claims of neutrality of human language are made, they are false: as we have argued, even the most mundane language is socially located and has a host of resonances.
Third, we doubt that neutrality is coherent as a speech ideal. Not only do we see no good reason why emotional appeals to, say, inequity in social location could not be components of healthy discussion, but we also see the purpose of discussion as intrinsically social. A theory of human discourse predicated on an ideal of neutrality, whether descriptive or prescriptive, would necessarily be incomplete, for it would omit the purpose of human discourse.
1. Reproduced in Wiesel, Night, 88.
2. Material from this chapter first appeared in Beaver and Stanley, “Neutrality,” and we gratefully acknowledge Matthew Congdon for very useful editorial comments as we prepared the journal version. Alice Crary provided a powerful response in her “Neutrality, Critique, and Social Visibility.” Although this chapter has been revised and expanded with material that connects it to the rest of the book, the central lines of argumentation are as in the original journal article, and we do not directly respond to her comments and criticisms in this revised version. Nonetheless, her response and conversations with her over the years have changed our thinking, and this has affected the final form of the chapter and the book as a whole. In particular, her comments regarding the importance of ideas in the post-Wittgensteinian ordinary-language tradition, and on certain strands of work within feminist philosophy, have led to us more clearly reflecting, at various points in the book, connections to work of Wittgenstein and Cavell, on the one hand, and work on standpoint epistemology, on the other.
3. Note that one might consider a weaker understanding of a “neutral space,” whereby while the space is neutral, the reasons expressed within it are not. Some might imagine a physical space such as a Roman forum as having this quality. We will not discuss this approach, save to note that we do not see that the mere fact of a forum being a physical space rather than an animate agent or a speech act makes it intrinsically neutral. A physical space embodies practices and comes to have resonances intrinsically tied to those practices. There was nothing neutral about a Roman forum, since only certain types of individuals would ever be heard there, and debate within the forum was firmly couched in terms of the pervading ethos. Physical spaces, like online spaces for debate, are not inherently any more neutral than a so-called neutral arbiter can be, for brief discussion of which, see below. We thank Leah Ransom for discussion of this issue.
4. Examples may be controversial, because part of the point of an emotional appeal is to supplant complacency and stir controversy, whether the goals are otherwise reasonable or not, and sometimes we may consider the emotional reactions to such appeals themselves to be in some sense reasonable, or at least appropriate. Consider the case of “abolitionist propaganda” discussed in Brooks, Bodies in Dissent. Henry “Box” Brown’s panoramas exhibiting the brutality of American slavery were meant to appeal directly to emotions, but the strong emotions they evoked were appropriate to the inhumanity of slavery.
5. Du Bois, “Of the Ruling of Men,” 83.
6. Butler, Excitable Speech, 86. Butler is commenting on MacKinnon, Only Words, and Langton, “Speech Acts and Unspeakable Acts.” 7. Here Butler is referring to a thread that runs through much of Habermas’s work, the idea that in rational communication, argumentation can play a role in the search for truth that transcends the particular perspectives and goals of interlocutors, and the interlocutors can be convinced not by each other, but by the force of the arguments themselves: “Argumentation can exploit the conflict between success-oriented competitors for the purpose of achieving consensus so long as the arguments are not reduced to mere means of influencing one another. In discourse what is called the force of the better argument is wholly unforced. Here convictions change internally via a process of rationally motivated attitude change” (Habermas, Moral Consciousness and Communicative Action, 160). Elsewhere, Habermas comments that interlocutors “are bound by presuppositions of communication and rules of argumentation that define ‘the space of reason.’ In this ‘space’ reasons can float freely and unfold their rationally motivating power unimpeded so as to affect the mind … in the right way” (Habermas, Truth and Justification, 140). The above quote from Butler makes the point that the concept of a deformation of speech presupposes that there is some sort of pure, undeformed speech, unsullied by ideology. Butler can be seen as standing in a philosophical tradition that critiques the possibility of reasons that transcend the socially embedded perspectives of reasoners. We also stand in this tradition.
Reasons cannot float freely.
8. Crary, “The Methodological Is Political,” 48.
9. “It is natural … to think of there being connected with a sign … besides that to which the sign refers … also what I should like to call the sense of the sign, wherein the mode of presentation is contained” (Frege, “On Sense and Reference,” 57).
10. “The sense of a proper name is grasped by everyone who is sufficiently familiar with the language” (Frege, “On Sense and Reference,” 57–58).
11. Frege, “Logic,” 140.
12. Frege, “Logic,” 140.
13. We are inspired here by Jennifer Foster’s discussion in “Busting the Ghost of Neutral Counterparts.” 14. The quote is from a posting by Livia Miller, “Writing Doggedly: Dog Idioms from around Europe,” on the Oxford Dictionaries OxfordWords Blog on July 11, 2013. At time of writing, this post is no longer available on the OxfordWords website, but is accessible at https://web.archive.org/web/20130718093339/http://blog.oxforddictionaries.com/2013/07/dog idioms-europe/.
15. Putnam, “Meaning and Reference.” 16. Grice, “Further Notes on Logic and Conversation,” 43. Grice is here discussing the difference between entailments and what he terms “conversational implicatures.” 17. Klemperer, The Language of the Third Reich, 20–21.
18. Klemperer, The Language of the Third Reich, 2–7.
19. Klemperer, The Language of the Third Reich, 2.
20. Balkin, Cultural Software, 102.
21. Perry, “Frege on Demonstratives.” 22. Frege, “The Thought: A Logical Inquiry.” 23.
Evans, “Understanding Demonstratives.” 24. Lewis, “Attitudes De Dicto and De Se.” 25.
Stalnaker, Our Knowledge of the Internal World, chapter 3.
26. Barthes, “The Discourse of History,” 132.
27. Statistics may be biased, but this is not in any straightforward way a case of hidden indexical perspective.
28. Moss summarizes this aspect of her account as follows: “Each de se proposition you believe is equivalent with some de dicto proposition, given what you believe with certainty.
This sort of de dicto proposition is something you convey to your audience, and something they come to believe.
Furthermore, your audience already has some de se beliefs about their relation to you. So they also come to believe some de se propositions: the consequences of their standing de se beliefs and their acquired de dicto information” (“Updating as Communication,” 235).
29. Stalnaker, Our Knowledge of the Internal World.
30. Egan expresses his conclusions about the inherently perspectival nature of meaning as follows: “attempts to quarantine de se content in the philosophy of mind, and maintain a pristine philosophy of language dealing only in de dicto contents, are doomed to fail. Once we let de se content into our philosophy of mind, there’s no viable way to keep it out of our philosophy of language altogether. It at least gets in to the pragmatics” (“De se Pragmatics,” 159).
31. MacFarlane, Assessment Sensitivity.
32. Classic papers include Lasersohn, “Context-Dependence”; Egan, “Disputing about Taste.” 33. Camp, “Imaginative Frames for Scientific Inquiry,” 308.
34. Camp, “Metaphor and That Certain ‘Je Ne Sais Quoi.’ ” 35. Camp, “Slurring Perspectives.” 36. Samia Hesni has made related arguments that intuitions about linguistic felicity are responsive to social location. She observes that contrasting intuitions as regards the felicity of sentences like “Muslims are terrorists” and “White men are terrorists” have as much do with the identity of the interlocutors as to do with facts in the world (“Normative Discourse and Social Negotiation,” 87–88).
Hesni’s argument is that if we wish to capture the evidential basis of the theory of meaning—speaker intuitions—we must theorize about social situatedness.
37. We recommend Rachel Fraser’s “Narrative Testimony” for explicit discussion of the relationship between narrativity and perspective, which she summarizes as follows: “In successful narrative testimonial exchange … speaker and hearer become co-ordinated not only with respect to the informational content of the speaker’s testimony, but also with respect to the content’s representational format. Such co-ordination gives rise to perspectival dependence” (4032).
38. It might be suggested that narrativeless discourse would be scientific. We will not contest whether there could be such a thing as narrativeless scientific discourse, but it is doubtful that actual scientific discourse is narrativeless. Here is Walter Fisher on the issue: “There is no question … that scientific discursive practices are rhetorical, whether involving the specialized audience of scientists or the generalized audience of the public. In both instances, rhetorical motives and rhetorical means are at work—the desire to gain adherence, adaptation to audience constraints, and the use of strategic, persuasive symbols, including charts, graphs, pictures, metaphors as well as empirical data and wellformed arguments. Justification to convince a specialized audience and translation for a generalized audience are both rhetorical ‘doings,’ practices” (“Narrative Rationality and the Logic of Scientific Discourse,” 22). In his earlier, and seminal, Human Communication as Narration, Fisher makes a related statement that bears on the supposed neutrality of argumentation in scientific discourse: “The most fundamental difference between narrative rationality and other rhetorical logics is the presumption that no form of discourse is privileged over others because its form is predominantly argumentative. No matter how strictly a case is argued—scientifically, philosophically, or legally—it will always be a story, an interpretation of some aspect of the world that is historically and culturally grounded and shaped by human personality. Even the most well-argued case will be informed by other individuated forms besides argument, especially by metaphor” (49).
39. Munton, “Prejudice as the Misattribution of Salience.” 40. Geuss, The Idea of a Critical Theory, 5.
41. Bourdieu, Outline of a Theory of Practice, 164.
42. Bourdieu, Outline of a Theory of Practice, 178.
43. Butler, Excitable Speech, 154.
44. In chapter 3 of Know How, Jason Stanley argues that first-person perspective does not pose distinctive issues for a broadly Fregean theory of content—it is just one kind of mode of presentation. Using similar considerations, Herman Cappelen and Joshua Dever argue in The Inessential Indexical that the focus on the specialness of de se thought to the theory of meaning has been misplaced.
45. Klemperer, The Language of the Third Reich, 14.
46. In the work of Elisabeth Camp, perspectives give rise to speech practices. One might disagree with the priority claim inherent in Camp’s work, though. Perhaps perspectives are also in part constituted by habits of behavior, including speech behavior. This, we take it, is what Klemperer is urging.
47. Sewell, “A Theory of Structure,” 8.
48. Paul, Transformative Experience.
49. Arendt, The Origins of Totalitarianism, 312.
50. See Lasersohn, Subjectivity and Perspective. The major work on Assessment Sensitivity is MacFarlane, Assessment Sensitivity. MacFarlane considers five applications, to taste, knowledge, statements about the future, and epistemic and deontic modals.
51. That is, the ubiquity of social perspective does not mean that social perspective must then play an analogous role in truth evaluation to the role played by parameters in MacFarlane’s account of predicates of personal taste (MacFarlane, Assessment Sensitivity).
For MacFarlane, this would undermine the defense he gives of a type of limited relativism, avoiding the pitfalls of taking all truth claims to be inherently unstable. He achieves this defense by separating those claims that depend upon parameters of evaluation from those that do not, an approach that would become untenable in the form he gives it were all truth claims relative to a parameter of social perspective.
52. We touched on Gibson’s work in section 1.5 and elsewhere in part I. The relevant reference is Gibson, “A Theory of Direct Visual Perception.” 53. Tarski, “The Semantic Conception of Truth and the Foundations of Semantics.”
CHAPTER EIGHT Straight Talk Confronted with an analysis … the code traffickers quickly turn into hard-core literalists who say things like “How can it be racist to be for equality?” or “My objection is simply to preferential treatment for anyone,” or “I just want decisions to be made on a basis that is fair.” Such demurrers invoke a plain meaning philosophy of language and also claim a transparency of intention: “I just mean what I say, and what I say is both innocent and upright.” The question is one of motives.
—STANLEY FISH1 One man’s transparency is another’s humiliation.
—GERRY ADAMS2 IN THE PRECEDING CHAPTER, we discussed and critiqued the ideal of neutrality of discourse. Since discourse is invariably shaped by social practices it presupposes and exemplifies, discourse is never neutral with regard to perspectival features such as (for example) social position. In this chapter, we discuss the limitations of another somewhat idealized conception of discourse—straight talk. Talking straight is about being transparent, keeping your communicative intentions up-front and obvious.
Neutral discourse has some intuitive and pretheoretic commonalities with straight talk. We can easily imagine the liberal ideal of discourse democracy to be a kind of neutral straight talk—discourse that lacks a perspective and is fully transparent. We saw in the last chapter that neutrality of discourse is incoherent, as perspective is essential to human discourse. An ideal that fundamentally misrepresents the nature of its domain is epistemologically problematic. The ideal of neutral discourse represents discourse as ideally free of perspectives. Neutrality as an ideal would lead one to think that transforming nonideal discourse into ideal discourse involves removing perspectives one by one until the ideal of neutrality is achieved. But discourse, as we have argued, is essentially perspectival.
Utterances of words characteristically exemplify speech practices, which locate their users socially. You cannot remove a perspective from a speech practice, except insofar as what you then have is a distinct speech practice, with its own perspective. Most ideals mask reality by making it seem that reality is closer to the ideal than it actually is. The ideal of neutral discourse masks the gap between it and reality in a more problematic way. Speech is fundamentally perspectival. If you remove speech practices, you lose speech, and speech practices are perspectival. The ideal of neutral discourse masks reality by completely misrepresenting the nature of the gap between it and reality, What about the ideal of straight talk? Here, we do not deny the bare coherence of straight talk. And perhaps it is logically possible to communicate fully transparently, though the ubiquity of speech practices makes that unlikely (as we shall see). Nevertheless, the ideal of straight talk is problematic in the theory of meaning, as it masks so much of what such a theory is responsible for explaining.
To understand how much of what this ideal masks, we turn to what we call hustle—the large body of talk exchanges that are not transparent, which constitute most or perhaps all of linguistic reality.
8.1. Defining Hustle via Straight Talk A communicative action is transparent, if the speaker thinks that the hearer will consciously recognize the communicative action as the action it is.
The straight talk in an utterance consists in those communicative acts the speaker performs transparently. Discourse is straight talk if it is composed solely of straight talk. The hustle in an utterance is all the communicative actions performed through that utterance that are not straight talk. In a nutshell, and to allow the use of hustle as a verb: Hustle is what people do with words nontransparently. They hustle. To see where this is going, let us note that hustle, thus defined, covers two major classes: (i) manipulative acts and (ii) communicative actions we perform but are not conscious of performing. First, suppose a speaker is manipulative.
For example, imagine us intentionally rushing you into signing a document without a careful reading of the fine print by suggesting that we have to leave in a few minutes. (“Busy, busy, busy!” one of us smiles, and looks down at an ostentatious watch.) We intend that you do not recognize the subterfuge; the speaker’s real goal in rushing you is to get you to overlook the fine print. Since we intend that you do not recognize it, and we also believe that you will not recognize it, the action of suggesting that we have to leave in a few minutes involved hustle (even if it happened to be true), namely convincing you to sign without reading the fine print.
Secondly, hustle includes unconsciously performed communicative acts.
Since we authors don’t know what actions we’re performing unconsciously, we’ll pick on a fictional, but plausible, third party as speaker.
Deep in the bowels of a British city, a female executive is carrying a stack of legal documents, and a male denizen of the office says, “Need some help with that, love?” The executive finds the offer, and the manner of asking, condescending, and replies as she strides past, “How about you go fetch me a coffee … black, no sugar. Thanks, LOVE!” The male worker knows he has been put in his place, but does not have a deep understanding of why or how. As he heads to the break room to get the coffee, he mumbles, shrugging, “I was only trying to help” to the small audience that witnessed the event. By offering help, the male worker suggested his addressee’s inadequacy for the task she was undertaking, and by referring to her as “love,” his utterance suggested, as far as she was concerned, a socially inappropriate level of familiarity between them, implying that she had equal or lower status.
At the time of utterance, the male worker did not think he was implying inadequacy or familiarity, and certainly wasn’t attempting to get his addressee to believe something that he himself did not.
Indeed, it is plausible to us that in such a case the male worker might not even consider the possibility that his words could be taken in such a negative light. If so, then it is clear that at the time of utterance he did not believe that his addressee would believe that he was performing the actions of suggesting inadequacy or familiarity. Therefore, these suggestions were hustle. Although there are practices that intrinsically involve hustle, such as lying and dog whistling, we have suggested ways in which hustle might accompany a much wider range of communicative actions, perhaps even all communicative actions. We have emphasized that hustle is central to the transmission of ideology and explicated mechanisms. Hustle has thus been central in this volume. Yet it is not one of the building blocks of our resonance-based framework: none of our core definitions of resonance, attunement, harmony, presupposition, and accommodation mention hustle.
The reverse is also true: hustle is not itself a resonance-theoretic notion. It is a general communication-theoretic notion that could be applied independently of the specific theoretical framework.
We can put it this way: hustle is about the failure of an idealization, the idealization of straight talk. The idealization is at the heart of models of meaning in a tradition of work developed by Paul Grice, which centers on the idea that for the speaker to mean something always involves a recognizable communicative intention. This is not to say that hustle cannot be modeled in frameworks based on recognition of communicative intention, but it would need to be modeled as an explicit divergence from what is taken to be ordinary communication. By contrast, the resonance framework was created to avoid strong commitments regarding the role of intention in communication, and without an idealization of straight talk. We no more control the resonances of the words we use than we control a mischievous genie, perhaps prepared to do our bidding after we release it from a bottle, but only in its own impish way.
The resonances of particular communicative actions derive from complex systems of practices, including practices of making sounds, practices of using words, practices of arranging the words, and rhetorical, interactional, and pragmatic practices. Although, as noted, certain practices that are of interest in theorizing the politics of language exist specifically because they can consistently be used to hustle, hustle applies at the level of the action, not at the level of the practice.3 Thus we can speak of the resonances of actions drawn from a practice being hustle on one occasion, and not hustle on another. Take pronouns, for example. The resonances of a pronoun such as “them” include joint salience of some group. That is: the use of the pronoun gives a probability boost to a certain feature of context, and that feature is the joint salience of some group. The relevant group will normally be manifest to interlocutors, and the intention to refer to it will commonly be obvious. But in section 4.4, we discussed cases in which the identity of the referent is not shared at all, and in which a pronoun is deployed strategically to give the illusion of such common ground. In such cases, the pronoun is not merely being used in a way that capitalizes on the presence of common ground. Rather, the use of the pronoun has disguised a function of creating or reinforcing common ground. The presence of such a strategic intention is not necessarily manifest to the hearer, and so the resonance of joint salience in such cases involves hustle.
Hustle in many of the cases discussed in chapter 4 was presumably intentional, but it need not be. Furthermore, hustle need not be a bad thing.
Creating common ground, for example, is usually seen as a good thing, and the fact that the mechanism behind the creation of common ground is not always overt does not thereby make it inherently problematic. Let us make some distinctions we can use in the remainder of the chapter. Hustle could be: 1. intended if the sender intends it not to be transparent; 2.
incidental if not intended; 3. unconscious if the speaker is not aware of performing the action; 4. detrimental if it is against the interests of the hearer; 5. malign if it is intended and detrimental, and malicious if the sender recognizes that malignance; and 6. congenial if at the time of the communicative act, the sender does not consider it to be detrimental hustle associated with that act.
Hustle is a broad and heterogenous class. It should be clear from this list that what is left out by the idealization of straight talk is a great deal of speech.
Let’s consider the Gricean approach mentioned above, where what something means hinges on what communicative intentions can be recognized. The assumption of recognizable intentions is even more central to Grice’s enterprise than the assumption of cooperativity (which we will discuss below). Without the assumption of cooperativity, Gricean reasoning doesn’t warrant pragmatic inferences, and pragmatic inferences are its trademark success. But without the assumption of joint recognition of intentions, not even literal content is conveyed. Loosely, according to Grice, an indicative utterance carries as part of its meaning a certain proposition if (a) the speaker intends that the audience will think that the speaker believes that proposition, and (b) the speaker intends that the speaker and hearer jointly recognize that the utterance is intended to yield that effect.4 This is what he calls nonnatural meaning, that is, the type of meaning achieved via communicative intention (as opposed to the smoke-means-fire type of natural meaning). One of the great virtues of Grice’s model of meaning is that it covers both literal meanings and the pragmatic implications that Grice termed implicatures. Let us illustrate the centrality of intention recognition to Grice’s model with a simple case of implicature.
Suppose that at some point Jason did not wish to go jogging with David.
He therefore developed an intention that David came to believe that it was a bad time for them to go out jogging. Jason said to David, “It’s raining.” David and Jason jointly recognized that Jason intended to achieve several effects by saying that. One was that David came to believe that Jason thought it was raining, that is, David came to believe that Jason believed the literal content of the sentence he had uttered. Another was that David came to believe that Jason believed that it was a bad time for them to go jogging, that is, David identified and accepted an implicature of Jason’s utterance.
As a result, David actually did come to believe that Jason believed it was a bad time for them to go jogging, and indeed that it was a bad time for them to go jogging, thereby fulfilling Jason’s wishes.
When Jason told David, “It’s raining,” was that straight talk? To a first approximation, where we only consider the communicated propositions that it was raining and that it was a bad idea to go jogging, yes. The action of communicating these propositions was certainly straight talk, since Jason believed David would recognize that he was communicating them, and we have considered no other effects. But when someone says something, there’s no end to the inferences that a hearer might draw. It just so happens that as a result of Jason’s utterance, and the sound of Jason’s voice as he produced it, David came to believe, among other things, that Jason had not fully recovered from his previous night’s revelry, and yet that Jason had not entirely lost the ability to communicate in English. Jason intended neither effect, and, in his somewhat diminished mental state, did not even have a belief that he would cause David to draw these inferences. So if we’re going to be strict about it, even Jason’s benign comment on the weather was not straight talk, although the additional hustle was both congenial and incidental. But note here that our definition of hustle is only as tight as its component terms, so that a generous conception of what Jason believed will lead to a correspondingly more restricted conception of which of the acts he performed were hustle.
While incidental hustle is, by definition, congenial, it may or may not be malign. It cannot be malicious. Intended hustle, on the other hand, could be congenial, malign, both congenial and malign, or not only malign but malicious. Suppose that on the morning before Sofia gives a big presentation, her boss says to her, “Don’t worry. Perhaps these guys aren’t as tough as everybody says!” Sofia didn’t have any prior indication that she’d be presenting to (what everybody says is) a tough crowd. The information was offered in the form of a presupposition, as if completely uncontroversial, and Sofia accommodates it readily. This causes Sofia to ramp up her presentation and give a really aggressive pitch. As it happens, the group she is presenting to does not like aggressive pitches, so that they sit stone silent after her presentation, causing Sophia great embarrassment.
It is immediately clear that she has lost the deal.
If Sophia’s boss did not intend for Sophia to decide to change her presentation, then causing that effect was incidental, but detrimental. If, on the other hand, her boss did intend that she would react to his comment by giving a particularly aggressive pitch, but thought that this would be beneficial, then her boss’s comment, qua action of realigning Sophia’s presentation style, was congenial but malign. If her boss not only intended to cause Sophia to make her pitch more aggressive, but also knew that this was not what was needed, then her boss was being actively malicious.
8.2. Coordination and Cooperation The notions of cooperativity and coordination will be important for us, so let us now consider how they relate to each other. First, as the terms are ordinarily used, coordination, unlike cooperation, does not imply benefit or intended benefit to others. Secondly, coordination does not require as high a degree of agency as does cooperation.
Coordination of activity between entities occurs when their behaviors are correlated in such a way as to make the behavior of one predictable given knowledge of the behavior of the others. Since coordination does not imply mutual benefit, one can coordinate with another’s actions in a purely self-serving manner. David might coordinate his trips to bars with trips of Jason’s, knowing that Jason will pay for expensive cocktails and without any desire to help Jason to reach personal objectives. We should not describe David’s actions as cooperative in this case, unless we take Jason as having a mutually recognized goal that is more easily reached as a result of this generosity.
Coordination does not require a high degree, or indeed, at the limit, any degree, of agency.
That is, coordination can occur without any intention to coordinate. Note first that we can coordinate our actions with the activity of inanimate entities. We can even coordinate our actions with nature itself.
For example, we might coordinate our gardening activities with the vicissitudes of the weather or coordinate our religious activities with the phases of the moon. Unless nature is itself agentive and benefits from our actions, these are not examples of cooperation. Note further that our actions may be coordinated with those of others even if there was never an intention to so coordinate. For example, in his little-known work as an undercover FBI agent, Jason was once forced to hide in the trunk of a mafia boss’s Lincoln Continental, a spot that a somnolent, or possibly recently deceased rival gang member by the name of Joey “Fangs” Napoli was already partially occupying, along with a set of slightly bloody golf clubs. Fortunately, the Continental has a large trunk.
For several hours, Jason’s movements were coordinated with those of the mafia boss.
Fangs’s movements too were coordinated with that of the mafia boss, as were the golf clubs’, although the latter two facts were true by design. Neither Jason nor Fangs was cooperating with the boss any more than the golf clubs were, and neither the boss nor the driver of the Continental had any desire to bring Jason along for the ride.
To cooperate, agents must choose to perform actions that enhance the probability that both they and others will reach their goals. To say that the activity of one entity is coordinated with that of another is to make a much weaker claim: it is to say that the activity of the first is dependent on that of the second. This dependency implies that knowing how the second acted would alter our best estimate of the probability of the first entity’s having acted as it did. To say that agents coordinate their actions (and not merely to describe their activity as coordinated) is to make a subtly stronger claim: it is to suggest that they choose actions in an interdependent way. On the strong idealizing assumption that agents are performing their actions as the result of rational choices, coordination implies that it is in their mutual interests to act in a coordinated fashion.
Since we will be using game-theoretic concepts below, it is helpful to relate our descriptions of cooperation and coordination to those standard concepts. Cooperative game theory concerns strategic problems in which it would be of interest for agents to use conventions to govern their activity, and even to subjugate their actions to an enforcing agency that verifies compliance with the conventions, thereby providing a guarantee of mutually beneficial coordination. Thus, cooperation involves coordination for mutual benefit, and conventions should be expected to emerge precisely when there is repeated occurrence of situations in which coordinated behavior for multiple parties is favored.
The point of current relevance is that what is understood as cooperative behavior, and thus the sort of behavior that cooperative game theory seeks to help us understand and regulate, is the sort of behavior a group might jointly opt to have externally enforced, because guaranteeing such action is maximally mutually beneficial. Take, for example, the most famous gametheoretic example of all, the Prisoner’s Dilemma. If two criminal accomplices recognize that in the future they might be in jail, and each is offered the chance to defect against the other, then it would be advantageous for them to set up a way to enforce nondefection. Mutually assured destruction, as we all know, can be a great motivator, even if it does little to help us sleep at night. Thus, the accomplices might agree that a third party is to observe their behavior, and if one of the two accomplices defects (by confessing to the police), then the defector is to be punished in some heinous way, for example, involving harm to their family. Having mutually agreed upon a mechanism guaranteeing such painful retribution, both parties will be confident of their cooperation with each other and not with the police, should they ever be apprehended. Thus, agreeing to possible retribution actually enhances their expected future payoff.
Now consider a communicative action that involves intended hustle. In order for there to be intended hustle, there must, by definition, be something believed not to be transparent to some second player, to whom the action must appear identical to another action that lacks any hustle. It follows, trivially, that there is such an alternative action: one just like the first, except lacking the hustle. This alternative action is the move that the second player takes the first to have made. It also follows that it would be problematic to assign payoffs such that the full set of players (a so-called grand coalition in game theory) would jointly agree to a regulation enforcing the first action rather than the second, for the simple reason that at least some players could not even distinguish the two actions. Therefore, communicative actions that involve intended hustle are not fully cooperative in the technical sense that they would not be accepted by a grand coalition in cooperative game theory.
This establishes that hustle is not fully cooperative in a technical sense, but we can go further: it is also the case that selfishness in action engenders hustle. If an action (or, more generally, a strategy, or, more generally still, a strategy profile for a subcoalition of agents) is selfishly motivated and is potentially suboptimal for other players, then that action has the following two properties. First, the type of action is not one that a coalition would choose as a norm in those circumstances, that is, the action is not suitable for conventionalization in the narrow sense of cooperative game theory.
Second, players taking such actions have a motive for hiding exactly what they did, keeping some of the desired effects of their actions secret. For it is only if their actions are disguised in this way that their defection can be expected to be unilateral, so that they make just the gains they anticipate, and the audience does not have the opportunity to reorient and redress some of those gains. This motive for secrecy becomes stronger when the reputation of senders with the audience is itself of value to the sender. It becomes stronger still when there is a societal norm governing the type of action performed and there is the possibility of onlookers observing or learning of the defection, thus damaging the sender’s reputation in society at large.
Summing up these considerations, we see both that lack of cooperativity in communicative behavior has a natural tendency to align with the use of intended hustle, and (our earlier conclusion) that communicative actions involving intended hustle are not fully cooperative.
Admittedly, the argument was formulated using a very particular and technical sense of what cooperation is, but the broad outlines of the argument are, we think, quite intuitive. It is even somewhat obvious, once one reflects on it, that people who are communicating in a noncooperative fashion might do well to hide their noncooperativity. Perhaps the claim itself is less remarkable than the fact that philosophers of language and linguists (including ourselves) have seldom reflected on the fairly intuitive relationship between nontransparency and noncooperativity, at least not in their lives as scholars.
Let us say that communication is coordinative to the extent that its effects enhance coordination, where we include here coordination of attitudes, emotions, and behaviors. For example, once upon a time, Jason screamed in terror, thus representing, although somewhat overeffusively, his emotional attitude toward a scary scene in a political documentary. As a result, David shared his terror. In this situation, coordination of emotional attitudes occurred, so in our terms there was, at the very least, a coordinative element to the screaming act.
The above arguments from cooperative game theory suggest that it may sometimes be useful to draw lines differently than our hustle terminology at first suggests, since talk involving hustle can be fully cooperative, provided it is unintended hustle. If we lump together straight talk and unintended hustle, we get what we will call prosocial communication, communication that lacks intended hustle. For example, since Jason did not have a hidden agenda, his screaming act was prosocial. A bee doing a dance is performing a communicative act, but is not being communicatively transparent, because bees, so far as we know, lack beliefs that their communicative actions will have effects. The communicative actions are all hustle. But it is presumably all incidental hustle. Now, a bee’s dance surely enhances coordination, which is a necessary condition for saying that the communicative act of dancing is cooperative. And since the bee also has no intention to be nontransparent, that is, all bee hustle is incidental, we can in fact say that in dancing the bee is prosocial, and in this respect is just like Jason screaming. Our readers may even feel it is apt to describe the bees themselves as cooperating. To describe them as such is to frame them as self-actuating intentional agents, whether that is intended literally or metaphorically.
As with bees, to describe a person as cooperative when they perform a socially significant act reflexively is to use an intentional framing, or to take an intentional stance, as Daniel Dennett would have it.5 Generally, we will find it easier to describe the person as cooperative the more deliberate and the less reflexive the action. Thus, a parent giving a distressed child a hug is cooperative, but someone screaming or sharply intaking breath is usually not aptly described as cooperating, even though their act may be prosocial in our technical sense, and fully cooperative in the sense that it enhances coordination and involves no intentionally hidden meaning. When Jason told David it was raining, which he did sincerely, he was certainly coordinative, since he (intentionally) enabled us to coordinate a complete absence of physical exercise for an entire day, and it was prosocial because there was no intended hustle. On the other hand, if Jason had told David the same thing in order to spare David the embarrassment of being outrun, that would have been coordinative but not completely prosocial. In such a situation, there is intended hustle in Jason’s persuasive actions, even though that hustle is congenial.
Our definitions lead to some subtle distinctions. Consider: David has bees, knows where a new source of pollen is, knows about bee communication, and has built a robotic bee. He remotely controls his robot bee to perform a dance for the colony, just as a regular bee would, indicating where the pollen is. As a result, the bees’ psychological states change, leading them to fly to the new source of pollen. Now, if a bee performed the same dance as David’s robot bee, that act would have been prosocial, and fully cooperative. But David, communicating via robot bee, had a communicative intention that he did not intend the recipients to recognize. Indeed, he knew that the recipients were unable to recognize communicative intentions. Like bee hustle, David’s hustle was congenial, but unlike bee hustle it was intended. Therefore, although the communicative act was in all relevant behavioral respects identical to the one that a bee would have performed, and although the bee’s actions would have been prosocial, David was not being fully prosocial. On the contrary, he was being manipulative. On this terminology, you are not being fully prosocial if, for example, you covertly manipulate your child.
8.3. Leading Questions Rhetoric, whether in the service of power or money, is clearly very successful in manipulating people. But how successful? Here is one result: asking people even purely hypothetical questions unconsciously shifts their subsequent preferences and behavior in often dramatic ways. In a study coming from a marketing perspective, by Gavan Fitzsimons and Baba Shiv, subjects were told in advance that they would be asked purely hypothetical questions.6 One group was asked, If strong evidence emerges from scientific studies suggesting that cakes, pastries, etc. are not nearly as bad for your health as they have often been portrayed to be, and may have some major health benefits, what would happen to your consumption of these items?
Subjects were told that the study “was about the effects of a change in environment on how consumers express opinions about products,” and so were directed into another room and offered a choice between snacks on a display cart placed between the rooms: chocolate cake or fruit salad.
Another group, the control group, was not asked any hypothetical questions.
In the control group, 25.7 percent chose cake. In stark contrast, subjects who were merely presented with the hypothetical question, and no further elaboration, selected the cake 48 percent of the time. Merely telling subjects to “please think carefully before you respond to the question” to prepare to justify their answer later increased cake selection from 48 percent to 66 percent. And subjects were clearly unaware of having been manipulated by the hypothetical question. Without exception, they denied that their preferences or their behavior were influenced by the hypothetical question in subsequent in-depth interviews. When asked whether they accepted that cake had major health benefits, every subject reported that they did not accept this and recognized that the question was hypothetical. And every single subject maintained that their choice was unaffected by being asked the hypothetical question.7 We already gave many examples, in chapter 4, of leading questions, although they were not hypotheticals, that is, not presented in the counterfactual mood. Let us consider a political example that is similar to the cases studied there, in that it contains a presupposition trigger of a type that has been widely discussed in the relevant semantic literature (the verb “know”), but is also similar to the unhealthy pastry case, in that it involves a hypothetical being used to shape preferences. Karl Rove was George W.
Bush Jr.’s campaign manager in the 2000 Republican primary that pitted Bush against John McCain. Before the South Carolina primary, Bush’s campaign polled prospective Republican primary voters with the following hypothetical question: “Would you be more likely or less likely to vote for John McCain if you knew he had fathered an illegitimate black child?” McCain had an adopted daughter from Bangladesh, and her skin tone is much darker than that of McCain or his wife. Presumably the question resonated strongly with certain voters.
Bush subsequently won South Carolina.
Clearly, hypothetical questions can be and are used as vehicles for intended hustle. Hearing hypothetical questions can change the way people think about an issue, priming an association so that when thinking about one thing, something else becomes salient. That is, the hustle can be thought of in terms of the power of the hypothetical question to coerce a change in the landscape of attunement, without conveying explicitly that certain propositions are true or false. This effect is not transparent to the hearer: that much is evidenced by the fact that hearers in the Fitzsimons and Shiv study denied any effect of the question.
A question can hustle whether the speaker intends such effects or not.
Suppose that a professor asks a student, “Have you read Adorno on this topic?” The professor might honestly intend this as a question of information, and yet the student might as a result (i) develop an intention to read Adorno, (ii) feel they should previously have read Adorno, and (iii) feel put down. If the professor does not intend these relatively predictable effects, then causing them is unintended hustle, whether or not the hearer actually recognizes them as effects of the utterance. In that case, the professor is prosocial but somewhat incompetent. Conversely, if the speaker does intend the effect, but does not intend that the hearer will recognize this intention, then that is also hustle, that is, intended hustle. One can imagine, if the professor’s communicative goal is satisfied, the student walking away from the conversation feeling somewhat down, without being able to identify exactly why. In that case, the problem with the professor is clearly not one of incompetence but temperament. Only if speaker and hearer mutually recognize that both the effects and the recognition of those effects were intended would causing the effects not be hustle. In that case they could be analyzed as (particularized) Gricean implicatures. The act of pragmatically implicating, at least in the sense originally introduced by Grice, is inherently prosocial, because it depends on transparency of intentions and a hearer’s ability to recognize those intentions. As Grice says, implicatures are calculable, meaning that a hearer can recover the reasoning that leads to the pragmatic inference. However, this prosociality seems to disappear if the use of implicature is motivated by a desire for plausible deniability that would have been unavailable had the relevant inference been stated explicitly. The professor saying “Have you read Adorno on this topic?” and intending for the hearer to recognize their inadequacy, can insouciantly deny that any such insult was intended.
Similarly, the professor could indirectly deny that any such insult could possibly have been intended by self-deprecatingly saying, for example, “I haven’t read him myself, but was just wondering.” On the other hand, the professor who goes on to say, “You haven’t read Adorno? You should be ashamed of yourself!” has made the insult explicit, and closed down any easy avenue for denying it.
If one effect of asking the question about familiarity with Adorno is that the student feels bad, then there is an element of malign hustle, and possibly malignant hustle if the professor intended the student to feel bad. Of course, it is impossible to say on the basis of a canned one-sentence example that the net effects would not be positive for the hypothetical student in the long term, but if the professor actually intended to cause the student to feel, for instance, marginalized and ignorant, even if only in the short-term, then we can say that there is a component of malignant hustle in the communicative act. Furthermore, even if the professor intended overall to benefit the student, the indirect approach of shaming them in order to spur them to greater future scholarship, if not a transparently stated strategy, would not be prosocial.
Rove’s push-polling question quite clearly involved malignant hustle.
Rove’s intention may have been that those taking the poll, as well as others deceived by the ensuing whispering campaign, would come to have false beliefs. Or maybe his intention was not to instill false beliefs, but rather to produce an emotional effect, to instill suspicion or concern. The development of a concomitant whispering campaign (of which this is a textbook case) would itself be intended hustle. Even if we generously ascribe to Rove the belief that a vote for Bush would be in the voters’ long-term interest, we take it that causing people to wrongfully suspect falsehoods to be true is inherently problematic. Perhaps this may be pardoned in some circumstances, perhaps in a scientific study such as the psychological experiments in which subjects were hustled by questions in chapter 4, or in the Fitzsimons and Shiv study. However, misleading people in a way that is designed to sway their voting behavior in a democratic election of national import is not excusable in the same way. We feel no need to varnish the claim that Rove and his team acted maliciously toward the voters. It seems reasonable to extend our terminology and say that when hustle is malign, it is not merely not prosocial, but simply antisocial. Karl Rove’s election tactics constitute a paradigm case.
8.4. Plausible Deniability The Gricean ideal of speaker meaning involves a communicative intention that the intention itself becomes mutual knowledge. This strong requirement of mutual knowledge is how the notion of overtness must be cashed out: an intention is fully overt if the speaker and hearer have mutual knowledge of that intention. When the speaker and hearer are regarded as having such a strong mutual epistemic commitment, denial is contradictory. As Richmond Thomason writes, Most conversational implicatures, I think, are meant; and in fact the response “I didn’t mean that” can always be used to renounce an implicature.8 Thus, if it is on the conversational record that a speaker has asked, “Are you busy tonight?” the speaker cannot deny having asked whether the hearer is busy without denying the conversational record itself. There is no room here for plausible denial.
Elisabeth Camp writes about the speech act of insinuation that “what is distinctive about insinuation” is that a speaker who insinuates is “prepared and able to coherently deny” the insinuated content.9 For example, if someone, after being stopped for speeding, utters “can we take care of this quickly?” they can deny offering a bribe when they are accused of so doing.
Camp labels this “implicature with deniability,” though we regard this as understating the degree to which the Gricean framework is challenged by this sort of phenomenon.10 The Gricean framework cannot make sense of the range of linguistic phenomena that allow for plausible deniability.
When there is an intention of plausible deniability, there is an absence of complete transparency about whether an inference was intended, and so, by our definitions, there is hustle: the talk is not entirely straight.
When Hans Gruber in the movie Die Hard says, “It’s a very nice suit, Mr. Takagi. It would be a shame to ruin it,” there is a clear suggestion that if Mr. Takagi does not cooperate, then the speaker will inflict serious bodily harm upon him, with collateral damage to the suit.11 In this case, where Gruber has previously asked for a code, and is holding a gun, we might naturally say that the implicated proposition is that if Takagi doesn’t give Gruber the code, Gruber will shoot. But the degree to which the suggested proposition can be fully specified is variable in such cases.
In a scene in the movie The Lincoln Lawyer, an ethically challenged client, Louis Roulet, who has broken into lawyer Mickey Haller’s home, says to him, “That’s a cute picture of your daughter, Hayley. She’s very pretty. She’s got soccer practice tomorrow, right?”12 There is a clear implication that if the Haller does not do Roulet’s bidding, his daughter will suffer.
13 In cases like this, the lack of mutual knowledge of what is implicated is evident: it’s in the nature of insinuation that it is unclear exactly what proposition is insinuated.14 Roulet’s is not merely a veiled threat, but a vague threat. It is difficult to identify with confidence a single, well-defined proposition that the speaker intends to convey. In this case, the criterion of mutual knowledge for an overt communicative intention is certainly not met.
Given the presence of a range of ill-defined propositions that are somewhat plausibly but not necessarily intended as implicatures, we are clearly dealing with intended hustle, although what we have is not complete opacity for the hearer, but rather semitransparency. Indeed, the hustle in these cases goes further than nontransparency as regards what propositions are intended. One wants to say that it is clear that he is not primarily complimenting Haller’s daughter and not primarily requesting information when he, superficially, asks about her soccer practice, but rather that the main action he is performing in uttering the entire three-sentence sequence is one of making a threat. But what threat is he making? If we cannot determine this, then we are not merely in a position of being unable to say what propositions Roulet is insinuating; we are also in a position of being unable to say, or even to know, in detail, exactly what Roulet is doing. Insinuation is a weapon of power, and the power is amplified by the ability to mask the details of the action, creating uncertainty and demonstrating that the insinuator holds all the cards.
The power relationship inherent to threatening insinuation is what is at play when it is subverted in Monty Python’s “Army Protection Racket” sketch, in which two wannabe mafiosi attempt to place untoward pressure on an army officer: DINO: How many tanks you got, Colonel?
COLONEL: About five hundred altogether.
LUIGI: Five hundred, eh?
DINO: You ought to be careful, Colonel.
COLONEL: We are careful, extremely careful.
DINO: ‘Cos things break, don’t they?
COLONEL: Break?
LUIGI: Well everything breaks, don’t it Colonel. [He breaks something on the desk.] Oh dear.… LUIGI: How many men you got here, Colonel?
COLONEL: Oh, er … seven thousand infantry, six hundred artillery, and er, two divisions of paratroops.
LUIGI: Paratroops, Dino.
DINO: Be a shame if someone was to set fire to them.
COLONEL: Set fire to them?
LUIGI: Fires happen, Colonel.
DINO: Things burn.15 Things do indeed burn. We leave readers to draw their own conclusions about the nature of brothers Luigi and Dino Vercotti’s hustle, and, for that matter, about the role of hustle in subversive humor.
To pick one more classic cinematic scene, consider the lengthy discussion between hitman Jules Winnfield and small-time student drugdealer Brett in the movie Pulp Fiction.
16 Jules’s palpable menace early in the scene (i.e., before it gets bloody) seems to stem not from any particular threatening speech act, but from his insouciance as he discusses breakfast and helps himself, with manifestly unnecessary urbanity, to Brett’s Big Kahuna burger. Jules’s casual demeanor shows that he will have no compunction in performing acts that transgress the bounds of normal society. It quickly becomes obvious that Jules can act as he wishes, and that Brett is totally in his power. But as the scene unfolds, neither Brett nor the audience can be quite sure what Jules is doing. Menacing insinuation is part of Jules’s modus operandi, a fact he makes explicit later in the movie, when he famously explains why he likes to dramatically recite a (partly fake) Bible quotation to his victims: “Now … I been sayin’ that shit for years.
And if you ever heard it, that meant your ass. You’d be dead right now. I never gave much thought to what it meant. I just thought it was a coldblooded thing to say to a motherfucker before I popped a cap in his ass.” If Jules just thought it was a cold-blooded thing to say to a motherfucker, and never gave much thought to what it meant, then it was a paradigmatic case of hustle. It was paradigmatic because there was no straight talk at all, and the function of Jules’s performance would presumably have been somewhat opaque to his victims, who, unfortunately for them, did not have a lot of time to think about it.
We would offer that perhaps the violence-prone among us are in the habit of making indirect threats and vaguely menacing insinuations for three distinct although related reasons. First, obviously, there is plausible deniability itself: it is hard to prosecute the Grubers, Roulets, Winnfields, or Vercottis of this world if they make their threats indirectly and with the possibility of denial. The second involves exploitation of the status of the proposition as plausibly deniable. While the mere fact of forcing the hearer to use their imagination to identify what actions the speaker has planned might add to the vividness of the threat, the suggestion that whatever is planned is so heinous that it cannot be talked about publicly amplifies the menace to a new level. The third is that the uncertainty raised by the lack of clarity of the intended message can itself be a source of distress either for a single addressee like the unfortunate Mr. Takagi, or for a larger audience among whom the uncertainty sows discord, both because they disagree about what was meant and how clear it was, and because they find it hard to know how to react to an individual who blatantly violates both norms of assertion, as well as norms of ethical behavior. This latter property of insinuation is crucial to Jules’s menacing behavior in Pulp Fiction. His control of the situation depends in part on his ability to appear to know exactly what he is doing, while nobody else does, and to be able to confirm or deny whatever he likes, while nobody else can.
We doubt that any of this reasoning was explicit for President Trump when, during the 2016 US election campaign, he famously commented of his opponent, “If she gets to pick her judges, nothing you can do, folks.
Although the Second Amendment people—maybe there is. I don’t know.”17 But it is clear that the speaker in this case understands very well that vagueness of a threat, and its deniability, can add to the menace. Indeed, the casual looseness with which an insinuation is dropped adds considerably to the effect, in part because it suggests that the speaker is not someone who can be reasoned with or trusted to stick within the narrow confines of traditional mores. The impossibility of reasoning is central to the so-called madman approach to diplomacy, a tactic that, within the United States at least, appeared until recently to have been left in the dust of Nixon’s fall from grace. It is also one aspect of the firehose of falsehood style of propaganda used in more recent times by Vladimir Putin’s regime (as discussed briefly in section 6.1).
By its nature, intended hustle requires cognitive sophistication: to hide something effectively, one must have a good understanding of the person one is hiding it from. Thus, if your best strategy involves lying or misleading, guaranteeing deniability, dog whistling so that different groups get different messages, or seeding the ground for a whispering campaign, you will be best off if you are working with a team of lawyers and PR specialists who either help you plan your move or clean up the mess afterward. The availability of such infrastructural support is one reason why politicians find it relatively easy to utilize intended hustle in their campaigns. Craven media is another.
To close this section, let us note a critical and perhaps underappreciated fact: there is a strong political asymmetry in the availability of hustle as a public rhetorical strategy. Veiled threats, insinuations, and simulated madness are all more easily and readily wielded by an authoritarian than by a democratic idealist, and all these devices quite obviously function rhetorically by way of intended hustle.
Here is a somewhat obvious premise: politicians advocating democratic ideals are thereby committed to a high degree of transparency. It follows not only that intended hustle should be inimical to them, but also that the costs of mounting a campaign involving intended hustle are potentially high. For if it is revealed that the democratic idealists are acting in a covert way, then they will immediately be revealed as hypocrites of the highest order. In contrast, a politician supporting authoritarian ideals has not thereby made any strong commitment to transparency. Therefore, the reputation of authoritarian politicians is not greatly endangered by the revelation that they have acted in a covert way. Indeed, it may even be a badge of pride.
The values of authoritarianism are centrally dominance and power; the goal is to win. Being ineffective will lose an authoritarian status; being opaque will not. The supporters of liberal ideals must be prosocial in order to live what they speak, or else expect to suffer electoral consequence. On the other hand, an authoritarian can not only survive the revelation that they are antisocial, but use this as positive evidence of their stop-at-nothing tactics.
The calculus is then that voters desperate for large-scale change will readily choose to have a stop-at-nothing strong man on their team, not merely in spite of antisocial infractions against democratic norms, but because of them.
8.5. Hustle and the Development of Speech Practice Whereas for intended hustle there is transparency for the speaker and opacity for the hearer, speech practices can lead to opacity for either speaker or hearer. In the former case, we have what we termed above unconscious hustle, a special case of incidental hustle. When you identify what seems to you to be, say, sexist language, but the speaker appears to be naively unaware of the overtones that so disturb you, that is incidental hustle.
Since mortal agents are incapable of having full access to the context and assumptions against which what they say or hear has meaning, everything that has been said or will be said by human beings will involve some degree of opacity for both speaker and audience.
The inevitability of incidental hustle means that we literally cannot have a full appreciation of what any sentence means. We cannot know all the resonances, any more than we can know the full history of the air we breathe. Furthermore, miscommunication, by its nature, involves incidental hustle. For, any case of miscommunication is a case where the requirement of straight talk is not met: if miscommunication has occurred, then there cannot possibly be mutual knowledge of the communicative intention to express whatever it is that was misunderstood.
Whether the inevitable opacity of meaning is a problem depends in part upon your goals. If your goal was to communicate just exactly one thing and nothing else, you would have a problem. But if your goal is, say, to increase coordination, then the fact that your utterances can have more meaning than you put into them is certainly not a problem per se. It may indeed be a benefit, the sort of thing a poet might strive for on occasion.
That is, it is at least possible to intend to achieve a certain effect on an audience, and to succeed in having that effect, and yet not to understand at all the mechanism by which your words produced that effect.
We submit that it is not only poets who might benefit from incidental hustle, but societies. We want to suggest, on the basis of mechanisms considered in the first two parts of the book, that incidental hustle can play an important role in the formation of group bonds and identities.
By engaging in a certain speech practice, one associated with a group to which the speaker or hearer both belong, say a common ethnic identity, interlocutors may increase coordination and collective attunement. As the community of practice develops, certain practices may become inwardly and outwardly meaningful to the practitioners (in the sense of section 6.4), without most of the practitioners recognizing that this process is occurring, and perhaps without anyone intending for it to occur.
Consider Canadian raising. This is a property of Canadian (and upper Midwest US) English affecting the sound of many words, so that, for example, “flout” sounds like “float” would in other varieties of English. It seems likely that the vast majority of productions contributing to the development of Canadian raising involved no conscious awareness of the ongoing sound change or what its exact rules were. People just did what their parents and acquaintances did, and for various reasons (like limited contact) felt no pressure to do what people across the border were doing.
That is, accommodation to the in-group by huge numbers of people, adapting their vowels to those of the people around them, led to collective harmonization around a shared practice.
Later, the distinctive practice may have become meaningful in and of itself, as a marker of identity, but it need not have developed as a conscious reaction to anyone else’s speech style, and indeed need not have developed with any strategic intent.
Though the example involves sound change, it should be clear that similar considerations might apply in cases of so-called semantic drift, whereby words change their meaning over time. This may happen in a community of practice of any size, whether a nation, a political movement, a cult, or a chatty pair of friends. Neither speaker nor hearer needs take stock consciously of the resonances their language has gained by being used previously within the group to which they belong. And yet, even without such awareness, the use of in-group language may improve communicative coordination among the interlocutors and contribute to cooperative success within the group more broadly.
8.6. Is Straight Talk Central?
By enumerating a range of applications, we hope we have set out a case that hustle is a worthy area of study in communication, and thereby that an idealization of straight talk is unreasonably limiting. One might protest that there are special foundational roles that are played by speech that fits into the straight-talk model. Is there, in the philosophy of language, a case for the “foundational nature” of straight talk? We discuss two arguments within philosophy of language for the view that speech that fits the idealizations of the standard model has a distinguished role. The first is David Lewis’s theory of metasemantics in his paper “Languages and Language.” The second is Paul Grice’s theory of implicature.
The project of semantic theory involves taking an assignment of semantic values to words in context and explaining how they combine to form the compositional semantic interpretation of the whole sentence, or perhaps the discourse. Semantics links a sentence like “dogs bark” to its literal interpretation, a tendency dogs have to make a certain noise. And the study of semantics does include difficult theorizing about the nature of the meanings, the semantic values, or words. For example, does “dogs” in “dogs bark” refer to a kind? Or is it rather a predicate of dogs, in which case perhaps the sentence “dogs bark” has an unpronounced generic quantifier that binds the subject of that predicate, as in “generically for x such that x is a dog, x barks”? Pragmatics is the study of aspects of the message conveyed by a linguistic act that are not part of the semantics. Metasemantics is a distinct project entirely from either semantics or pragmatics. Metasemantics is the project of explaining (roughly, because the word “meaning” is here used pretheoretically) why linguistic expressions have the meanings they do. Lewis’s “Languages and Language” is a contribution to the project of metasemantics.
Lewis’s distinctive take on how to proceed in metasemantics requires preliminary explanation. Lewis takes languages to be mathematically defined abstract entities, functions from words (and larger expressions, such as sentences) to meanings (or “semantic values”).
Languages are defined by recursive definitions of truth, as in the classical Tarskian definition of truth in terms of satisfaction. This sets up the problem of Lewis’s paper, which is to describe the relation between groups of people, or populations, and these abstract languages. The relation Lewis describes in “Languages and Language” is what yields his answer to the question “Why does this population speak this language?” and hence derivatively, the answer to the metasemantic question of why expressions have the semantic values they do.
So the goal of David Lewis’s “Languages and Language” is to describe what makes it the case that a given population P speaks a given language L, where languages are here conceived as functions from syntactic strings to semantic values, “semantic systems discussed in complete abstraction from human affairs.”18 Lewis’s answer is that the population P uses the language L if and only if “there is a convention of truth and trust in L.
To be truthful in L is to act in a certain way: to try never to utter sentences of L unless one believes it true in L. To be trusting in L is to form beliefs in a certain way: to impute truthfulness in L to others, and thus tend to respond to another’s utterance of any sentence of L by coming to believe that the uttered sentence is true in L.”19 If Lewis’s analysis of what it is for a population P to use a language L is correct, then “straight talk” has a foundational metasemantic role. It is when we engage in straight talk, using language to say what we mean and mean what we say, that we engage in the kind of use that underlies the truism “meaning is determined by use.” On this account, straight talk, where we are speaking truthfully and seriously, determines what language we are using. Since the language we are using is a function from symbols to meanings, on Lewis’s picture, straight talk is a central element in the metasemantic story explaining why we mean what we do by our words.
Hustle is not.
In “Coordinating with Language,” Jessica Keiser convincingly responds to Lewis.20 Here is an objection Lewis considers toward the end of “Languages and Language.”21 There could be some people, let’s call them “New Yorkers,” who use a language, but whose typical use of that language involves sarcasm, innuendo, and other forms of speech that is not what one might think of as straight talk. Based on consideration of these hypothetically hustling New Yorkers, Lewis argues that we need to characterize a notion of serious communicative speech situation, and he claims that it is uses of language in serious communicative speech situations that determine whether or not that population uses a particular language, abstractly conceived. There is much work to be done to justify Lewis’s claim here.22 A serious communicative situation, for Lewis, is a situation that “exists with respect to a sentence S of L whenever it is true, and common knowledge between a speaker and a hearer, that (a) the speaker does, and the hearer does not, know whether S is true in L; (b) the hearer wants to know; (c) the speaker wants the hearer to know; and (d) neither the speaker nor the hearer has other (comparably strong) desires as to whether or not the speaker utters S.”23 Lewis’s proposal is that a population P uses a language L if and only if there is a convention of truthfulness and trust in L in serious communication situations. But why think that the only uses of language that are meaning-determining are ones that occur in serious communicative situations?
Using the sentence “Hannah is in New York” in lying to someone is not a serious communicative situation. Nevertheless, someone using “Hannah is in New York” to lie is still using the words with their meanings in L. Why think that these uses are not meaning determinative? If I sarcastically say, “Trump will be a great president,” my use of the name “Trump” still denotes Trump. Political discourse is not generally made up of serious speech situations. But words are nevertheless used with the meanings they have in the language.
Mutatis mutandis for sentences uttered in other nonserious speech situations. All that use somehow fixes the fact that population P uses language L.24 Why think that there is any story about the connection between use and meaning simpler than this? Keiser recommends that we think of it another way. Perhaps our New Yorkers are always sarcastic, always joking, always intending to amuse at the same time as inform. Lewis’s view entails that they do not then speak a language. That consequence is just false. We can imagine New Yorkers as described, and they would speak a language.
Perhaps the claim is conditional; it’s enough to establish a convention if members of a population were to be in serious communicative situations with respect to sentences of L, they would have a convention of truth and trustfulness in L in those situations. But why privilege serious communicative situations? Couldn’t there be semantic properties of language that only are brought out by uses of sentences to do more than convey information in a jointly cooperative project? Would an analogous claim about phonological and grammatical properties, that they are all revealed in serious communicative situations, be at all plausible?
There is no independent reason to think that all the semantic properties of a language are brought out by straightforward information exchanges.
Many people do not find straightforward information exchanges particularly natural means of communication, and it’s just not at all clear why we should think such environments are special in that the kind of use in these situations completely determines the language a population is using (or, in perhaps more natural terms, completely determines all the semantic properties of words).
Meaning is determined by use, to be sure. But there is no reason at all to think there is any kind of magical formula for determining which uses are meaning determining, no reason to think that we can characterize in advance a class of uses that has a special, exhaustive role in the determination of meaning. And so there is no reason to accept Lewis’s characterization of the central class of cases that do all the work of metasemantics, and hence no case for privileging speech that fits the standard model.
That said, even if Lewis’s metasemantic theory were correct for some portion of language, what might be called descriptive language if it were cleanly separable from the rest of language, it is irrelevant to our project in this book. That fact would not justify the use of the standard model as a guide to what is of overall importance in the theory of meaning. If we seek an explanation of how linguistic communication functions, we must have a model that treats hustle and straight talk on a par. This is consistent with an account that assigns a special metasemantic role to straight talk for just some aspects of language. The goal of metasemantics on the Lewisian picture is to state what makes it the case that a population uses its words to mean what they do, that is, speaks the language that they do. Speech that fits the standard model may have a special metasemantic interest for some, say those studying the theory of reference. However, that should not lead us to privilege the model in a different task, giving a theory of linguistic action. A theory of linguistic action must extend beyond actions of description.
Are there other reasons to give a special role to transparent communication? In Grice’s classic work on implicature, he analyzes implicatures, a class of messages delivered by utterances that are typically not semantically encoded by their words and modes of combination. He relies upon one of the idealizations we criticize: that, in general, speech is a cooperative enterprise. One might take this as evidence that cooperation is the right model to assume, as it explains even a good deal of apparently noncooperative speech.
Here are two related rebuttals of the claim that Grice’s analysis of implicature essentially relies on constraining attention to cooperative contexts. The first is from Jennifer Saul.25 If you tell someone who’s expecting you to do some errands “I’m on my way!” while in fact you’re at home, you could reasonably be said to have lied. It matters not that you (claim to) intend to mean that you’re “on your way” to fame and fortune, since that’s not a contextually appropriate completion. Saul suggests you have lied if you (correctly) believe all contextually appropriate completions to be false. She is arguing that a central notion of Grice’s account, namely what is said, must be analyzed in a quite specific way to account for the difference between lying and misleading. This implies that considering noncooperative contexts can shed light on notions fundamental to Grice’s account, and that ignoring them may lead to incompleteness. Underscoring this point, “Conversational implicatures are often exploited as a way to mislead without lying.”26 Here is a second and somewhat more direct reason to think that an account of implicature cannot rely on an assumption of cooperation, although it is closely related to the version of Saul’s argument above.
Whereas the first argument shows that in cases where interlocutors might incorrectly assume full cooperation, implicatures are still generated, we now turn to adversarial cases.
Here it is manifest that interlocutors cannot make any strong assumptions about cooperativity. In Nicholas Asher and Alex Lascarides’s 2013 paper “Strategic Conversation,” they point out that implicatures are generated even in clearly noncooperative speech situations, such as clearly antagonistic courtroom cross-examinations.27 Some of Saul’s key examples of misleading speech are courtroom cases as well. By focusing on cases where there is no basis for strong assumptions of cooperativity, Asher and Lascarides conclude that we need a more general theory to explain implicatures, one that does not accord Gricean cooperativity any special status. One might also grant that strong cooperative principles like Grice’s principle of cooperation explain some features of speech, including many classic cases of implicature. But that is far from admitting that cooperative speech situations have a foundational role. Hustle is vastly more varied than the study of implicature.
To generalize from these two arguments using the terms we introduced in section 8.2, we must still be able to identify what is said in the case of utterances that are not prosocial.
28 We sketched there a general argument for the view that non-prosocial actions, that is, communicative actions involving intended hustle, are not fully cooperative. We do not wish to claim that it is an easy project to identify what is said when people hustle, merely that a theory that ruled out the possibility of identifying what is said in such cases would be of extremely limited interest in a theory of communication. Indeed, we should prefer a theory of meaning that accounts for what is said in overtly antagonistic situations that might arise in a courtroom or in an election campaign, and even for the antisocial utterances deployed by political manipulators like Karl Rove in his push polling operations.
It should be clear that hustle is a category considerably more expansive than just lying. This should be obvious, because unintended hustle is not lying. But is intended hustle simply lying? Lying involves an (intended) mismatch of belief between speaker and hearer as regards the proposition the speaker intends to communicate. It’s useful to show in detail that such a mismatch is neither a necessary nor a sufficient condition even for intended hustle.
Consider the example of Linda and Sofia, who will celebrate their anniversary on Friday, that being, coincidentally, their regular weekly date night. Linda wants to go to a certain restaurant to celebrate, L’Auberge du Philosophe, and wants Sofia to believe that she would really enjoy eating there on their anniversary. They have not discussed their anniversary, and Sofia has in fact forgotten about it. On Tuesday Linda mentions that a colleague of hers clearly earns too much, as he goes to L’Auberge du Philosophe twice a week, and insists on regaling her with mouth-watering tales of their most special concoctions. Apparently, the menu du jour is always fantastic. Linda appears blissfully transported to another place as she reports on these tales. On Wednesday, Linda is reading, and looks up and smiles at Sofia. Sofia asks Linda what she is reading, and she replies, truthfully, “It’s a review of L’Auberge. Five stars!” Later that day, Sofia impulsively calls L’Auberge, thinking it’s been awhile since they treated themselves to a really special meal. She is fortunate enough to find that there’s a free table that very Friday. Suddenly, she has a further thought … August 12, what is special about that day? “It’s our anniversary!” she blurts out to the receptionist. This leads to a somewhat more personal conversation than she had intended, but the receptionist is demure throughout. Blushing a little as she puts down the phone, she thinks, Good catch, Linda will really enjoy eating at L’Auberge on our anniversary! I’m so glad I thought of taking her there!
In the story of Linda and Sofia’s anniversary meal, Linda succeeds in causing Sofia to believe a certain proposition, namely that Linda would enjoy eating at L’Auberge on their anniversary. Because of Sofia’s apparent cluelessness, with which the authors cannot help but identify, Linda’s plan could easily have failed. It is nonetheless the case that Linda intentionally caused Sofia to believe the proposition in question by virtue of producing two utterances. That proposition, furthermore, is one that Linda herself believed. It is also the case that what Linda said was literally true. Yet we would describe this as congenial hustle on Linda’s part. It is congenial because the speaker intends that the hearer comes to believe something that the speaker thinks is true, and the speaker thinks that it is in the hearer’s interest to believe it. Yet it is hustle because the speaker did not think that the hearer recognizes the speaker’s true communicative action. Thus Linda’s actions are not fully prosocial, but to the extent that the mistaken beliefs Sofia formed are not harmful, Linda is not being actively antisocial.
Congenial hustle is common in conversations with children. If we tell a child that going on a hike in the park would be fun, we might receive in return a contradictory response, followed rapidly by a hardening of the child’s attitude. For this reason, a caregiver might avoid what psychologists would term the central route to persuasion, which operates by presenting the target with facts and arguments (as introduced in section 2.6). Instead, we might mention either to the child, or to another addressee within earshot of the child, how good the ice cream is at that little place just down the road from the park entrance. Psychologists of persuasion call this a peripheral route.29 In following this route, we do not intend that the child recognizes our communicative intention. Of course, we have only the child’s best interests at heart, and mean no harm by our communicative sleights of hand (which is not to say that they will not come back to haunt us as the child gets wiser, and we do not).
The anniversary and ice cream examples illustrate the possibility of hustle without dishonesty on the part of the speaker as regards the literal content of the utterance. The speaker is not trying to get the child to believe something false; what we have here is hustle, but it is not lying. There is no intended mismatch here between speaker and hearer on whether the literal meaning is true. There is a mismatch as regards the speaker’s underlying intention, but for child-directed speech it is far from clear how seriously we could take the view that understanding usually involves intention recognition in the first place. The prevailing view is that theory of mind (which is understood in psychology as a capacity, rather than as a scientific theory) is a developing trait in young children. This implies that intention recognition emerges gradually rather than being central to all child communication.
Next, we illustrate the possibility that speakers might intend that a proposition they do not themselves believe will come to be a belief of the hearer by virtue of a communicative act.
Suppose Sofia thinks it will rain, but Linda doesn’t. They discuss this and agree to disagree. Linda knows the umbrella is in the closet, but Sofia doesn’t know where it is. Linda says, “The umbrella is in the closet.” Sofia comes to believe that if she goes to the closet and fetches the umbrella, it will lower the chances of getting wet.
Linda intends Sofia to believe that, and Sofia knows that she does.
However, Linda does not believe it herself, since she believes that the probability of rain is zero. She has been cooperative. (One might even say Linda is displaying supererogatory communicative behavior. She is not merely cooperating, but playing along with what she sees as a misguided worldview to keep Sofia happy.) Question: is Linda hustling? Intuitively, no, because she is being, at least as regards the issues of immediate relevance, completely transparent. This exemplifies the nonsufficiency of the mismatch condition. Hustle is not about mismatch per se, although of course mismatch of belief is a necessary precursor not only to hustle but to communication itself; individuals with identical beliefs have limited need to exchange information about the world. No, hustle is not about mismatch.
It’s about transparency.
In considering the ice cream case, the umbrella case, or, for that matter, Saul’s Lying, Misleading, and What Is Said (which has extensive moral dimensions beyond what we have discussed), we of course are aware that lying and misleading are topics of perennial philosophical and political concern. So far as we can see, the developments in this volume do not bear on Kant’s famous arguments that lying is morally indefensible. But what of the even stronger claim one might make that hustle is morally indefensible, and thus that all talk should be straight talk? We must certainly deny any such claim, but not on moral grounds.
We must deny it because it ignores the inevitability of hustle in human communicative practice.
What, then, of the ideal of straight talk? Most speech hustles. Adopting the idealization of straight talk would make hustle fall outside the ambit of the theory of meaning. We rejected the idealization of neutrality because it is incoherent. Straight talk is not entirely incoherent.
Straight talk might be a useful political concept, or a guiding ideal for discourse in an ideal language, say for mathematical or logical inquiry. The straighter your mathematics, the better, it might be claimed. But as an ideal guiding model building for a theory of meaning for natural language, it should be rejected.
Building concepts and tools drawn only from consideration of data that is straight talk will result in a partial theory at best. In parts I and II of this book, we developed concepts and tools for a novel theory of meaning. These concepts and tools—and the theory—were straightforwardly based on speech that was hustle, which compromised the core of our data throughout. Even the “oops,” “ouch,” and nodding behaviors of chapter 1 were analyzed in practice-theoretic terms that, despite being inherently interactional, involve no reference to communicative intention.
If one centers hustle, one is immediately led to recognize the centrality of speech practice to a theory of meaning. The ideal of straight talk masks the importance of speech practice. We have no proof that some of the tools and concepts of our theory of meaning, like the concepts of resonance and attunement, could not be derived from consideration of straight talk alone.
But this is not how we derived them and assigned importance to them. We have no a priori argument against straight talk as an idealization, but in developing our theory, we have totally ignored it.
1. Fish, There’s No Such Thing as Free Speech, 90. 2. Adams is cited in “N.Ireland Process: Where Did It Go Wrong?,” CNN, October 22, 2003, http://www.cnn.com/2003/WORLD/europe/10/22/n.ireland.hold/. The context was negotiations over the IRA’s decommissioning of arms. In the above quote, Adams, the leader of the political party associated with the IRA, Sinn Fein, was indicating that full disclosure of exactly what arms the IRA had in its possession and was decommissioning would be humiliating to the movement.
3. This is not to say that the term “hustle” could not be defined at the level of practice. There are communicative practices for which a primary function is to hustle, for example boasting, lying, or dog whistling, and so one might reasonably talk of these as “hustling practices,” or some such.
4. Grice, “Meaning,” 383–84.
5. Dennett, The Intentional Stance. For evidence that treating an entity in a joint task involves a substantively different mental representation of that entity, see Gallagher et al., “Imaging the Intentional Stance.” 6. Fitzsimons and Shiv, “Nonconscious and Contaminative Effects of Hypothetical Questions.” 7. Fitzsimons and Shiv, “Nonconscious and Contaminative Effects of Hypothetical Questions,” 234.
8. Thomason, “Accommodation, Meaning, and Implicature,” 345.
9. Camp, “Insinuation, Common Ground, and the Conversational Record,” 45.
10. Camp, “Insinuation, Common Ground, and the Conversational Record,” 46.
11. McTiernan, Die Hard.
12. Furman, The Lincoln Lawyer.
13. In the conversation in The Lincoln Lawyer, the tables are soon turned: Haller: Don’t.
Roulet: Don’t what?
Haller: You think you’re the first client to ever threaten me with my family?
Roulet: Huh? All I said was she’s pretty.
Haller: Are you scared, Louis? Because where you are, right now … you’re in a very dangerous place. 14. See Fricker, “Stating and Insinuating,” who makes this central to her account.
15. Chapman et al., Monty Python’s Flying Circus, 97–98.
16. Tarantino, Pulp Fiction.
17. Trump is quoted in Corasaniti and Haberman, “Donald Trump Suggests ‘Second Amendment People’ Could Act against Hillary Clinton.” 18. Lewis, “Languages and Language,” 166.
19. Lewis, “Languages and Language,” 167.
20. Keiser, “Coordinating with Language.” One objection we can sketch but do not defend is related to “Foster’s Problem” for truth-conditional semantics. The theory of meaning Lewis employs, possible-worlds semantics, tends to underdetermine meaning. “Hesperus is a planet” and “Phosphorus is a planet” have the same truth conditions but different meanings.
Even more problematically, the predicates “is red” and “is red and Peano Arithmetic is incomplete” have the same possible worlds truth conditions, but clearly have different meanings. A convention of truth will help link populations to languages in Lewis’s sense, with possible-worlds semantics values, but is less obviously helpful with more realistic theories of meaning that individuate content more narrowly than truth conditions.
21. “Suppose they are often untruthful in L because they are not communicating at all. They are joking, telling tall tales, or telling white lies as a matter of social ritual. In these situations, there is neither truthfulness nor trust in L. Indeed, it is common knowledge that there is not” (Lewis, Languages and Language, 183). 22. Jessica Keiser (Non-Ideal Foundations of Language, chapter 2, 17–41) persuasively advances these and other theoretical objections to Lewis’s metasemantic picture. She goes on to provide a detailed alternative version that does not accord straight talk a foundational metasemantic role.
23. Lewis, “Languages and Language,” 183.
24. Or speaking in a framework we find more natural than Lewis’s, these facts fix the meaning conventions for the words of the language, now considered as a symbol system independently of an interpretation.
25. Saul, Lying.
26. Saul, Lying, 76.
27. Asher and Lascarides, “Strategic Conversation,” 15–16.
28. We thank Jennifer Saul for discussion of ways in which hustle might be a considerably broader category than what she calls “misleading” in Lying.
29. Petty, Goldman, and Cacioppo, “Personal Involvement as a Determinant of Argument-Based Persuasion.”
CHAPTER NINE Philosophy and Ideal Theory In philosophy we often compare the use of words with games and calculi which have fixed rules, but cannot say that someone who is using language must be playing such a game.
But if you say that our languages only approximate to such calculi you are standing on the very brink of a misunderstanding. For then it may look as if what we were talking about were an ideal language. As if our logic were, so to speak, a logic for a vacuum. Whereas logic does not treat of language—or of thought—in the sense in which a natural science treats of a natural phenomenon, and the most that can be said is that we construct ideal languages.
But here the word “ideal” is liable to mislead, for it sounds as if these languages were better, more perfect, than our everyday language; and as if it took the logician to shew people at last what a proper sentence looked like.
—LUDWIG WITTGENSTEIN1 WE HAVE THUS FAR EXPLORED two idealizations at length: neutrality and straight talk.
We argued at length against both. We begin this chapter by situating our project first within the broader ambit of traditions across philosophy that critique idealizations. We precisify the target of ordinarylanguage philosophy—the idealizations of “ideal-language theory.” We conclude by showing that the theory we have developed in part I of this book, centering on resonance and attunement, demonstrates that ordinarylanguage objections to these idealizations about language cannot be rescued by standard defenses of ideals in epistemology.
2 9.1. Nonideal Epistemology and Beyond When economists, philosophers, and political scientists study rationality, they commonly take as a model an agent with an infinite amount of time to reason and perfect logicality. This perfectly rational agent has utilities ordered linearly, to allow their preferences to be thought of as a utility function that orders choices along a dimension that can easily be ranked, for example, as an amount of money as measured in dollars. They use their knowledge to flawlessly perform judgments, and combine that with their preferences to make irrefutably optimal decisions. Their actions conform perfectly to the results of these decisions. The attitudes and dispositions of the rational agent are in perfect harmony. It is quite clear that individual persons are not the perfect rational agents of the study of rationality. We do not have an infinite amount of time to make decisions. As a consequence, we simply cannot consider every possibility, as this would take infinite time. It is even clearer that we lack perfect logicality; we are prone to errors in reasoning and not just when we are tired. How relevant to understanding humans are the properties of perfectly rational agents with no memory limitations? How relevant are the properties of such models to the study of anything of genuine scientific concern?
There have been a number of challenges over the past fifty years to standard models of rationality.
3 The standard model of rationality emerges from the study of the mathematical properties of idealized agents. But one might reasonably adopt another approach, one that begins with actual agents and the limitations they face. In philosophy, this is the naturalist project in epistemology. Naturalizing epistemology requires us to think of the agent that we study as located in space and time and limited accordingly.
Naturalism leads philosophers to consider the constraints due to the physical location of the knower. But the physical location of the knower is also a social location. The social location of a speaker is what is relevant to understanding whether they have a standpoint of special authority over their subject matter. To take an obvious example: if someone has a disability, they have a social location with special authority over the obstacles facing someone with that disability. Someone who is hearing impaired generally knows what it is to be hearing impaired better than someone who is not hearing impaired, as both authors of this volume can attest, having each experienced both functional and impaired hearing.
Naturalism in epistemology and philosophy of science made it difficult if not impossible to ignore the “situated knower” of feminist epistemology, critical race theory, and Marxism.
Charles Mills writes, Though mainstream philosophy and analytic epistemology continued to develop in splendid isolation for many decades, W. V. Quine’s naturalizing of epistemology would initiate a sequence of events with unsuspectedly subversive long-term theoretical repercussions for the field.… If articulating the norms for ideal cognition required taking into account (in some way) the practices of actual cognition, if the prescriptive needed to pay attention (in some way) to the descriptive, then on what principled basis could cognitive realities of a supra-individual kind continue to be excluded from the ambit of epistemology?
For it then meant that the cognitive agent needed to be located in her specificity—as a member of certain social groups, within a given social milieu, in a society at a particular time period.4 In her Stanford Encyclopedia of Philosophy article “Feminist Epistemology and Philosophy of Science,” Elizabeth Anderson raises a concern about idealizations in much of epistemology. Anderson is concerned about the narrowness of scope of the evidence that analytic epistemologists have used as the basis of their theorizing. She argues that this artificially makes certain knowledge claims paradigmatic, and others unusual, difficult cases to explain. But this is due merely to the narrowness of the idealizations, not epistemic reality: Mainstream epistemology takes as paradigms of knowledge simple propositional knowledge about matters in principle equally accessible to anyone with basic cognitive and sensory apparatus: “2 + 2=4”; “grass is green”; “water quenches thirst.” Feminist epistemology does not claim that such knowledge is gendered.
Examination of such examples is not particularly helpful for answering the epistemological problems that arise specifically in feminist theory and practice. What is it to know that I am a woman? What is it like to be sexually objectified? Why is it that men and women so often have dramatically divergent understandings of what happened in their sexual encounters? How can we arrange scientific practices so that science and technology serve women’s interests?
These kinds of questions make other kinds of knowledge salient for feminist epistemology: phenomenological knowledge, de se knowledge, knowledge of persons, know-how, moral knowledge, knowledge informed by emotions, attitudes, and interests. These kinds of knowledge are often gendered, and they can influence the propositional claims people are disposed to form and accept. This has critical implications for mainstream epistemological conceptions of knowledge, insofar as the latter are based on false generalizations drawing only from examples of ungendered knowledge.5 Feminist epistemology has been widely taken to be successful in vindicating Anderson’s concern about the traditional examples of analytic epistemology. When one looks more broadly at epistemology, including political examples, previously mysterious cases—such as the de se, or knowing how—become less so. It may have appeared that the content of basic knowledge claims was timeless, aperspectival, and value-neutral. But if so, that only shows that the notion of a “basic knowledge claim” was playing problematic ideological work in restricting examples. We have precisely similar concerns to the ones Anderson here gives voice about the theory of meaning.
Our specific understanding of the issues has been shaped by Charles Mills’s 2005 paper, “ ‘Ideal Theory’ as Ideology,” in which Mills calls attention to the dangers of an analogous situation in political philosophy, the privileging of a methodology of studying cases under certain liberal idealizations.6 One example is Rawls’s restriction to “well-ordered societies,” which in A Theory of Justice he sees as “cooperative venture[s] for mutual advantage.”7 Mills calls this methodology “ideal theory.” Despite the label, Mills’s concern with ideal theory is not that it employs ideals.
Mills is clear that any normative theory employs ideals. His critique is rather of the view that the primary, or the “foundational,” project of political philosophy is to construct a theory of justice for institutions and states under certain particular idealizations, idealizations that mask or marginalize many of its central topics. Consider the concept of social change. To study social change, one needs to theorize from a broader class, one that includes both societies that are not “well-ordered” in Rawls’s sense and ones that are. To consider as “foundational” only the questions that can be posed by restricting one’s attention to societies and institutions that fit the ideal model suggests that the question of social change is not as fundamental of a problem.
Perhaps Mills’s greatest concern is that the methodology of ideal theory will lead to “silence on oppression”: Almost by definition, it follows from the focus of ideal theory that little or nothing will be said on actual historic oppression and its legacy in the present, or current ongoing oppression, though these may be gestured at in a vague or promissory way (as something to be dealt with later). Correspondingly, the ways in which systematic oppression is likely to shape the basic social institutions (as well as the humans in those institutions) will not be part of the theory’s concern, and this will manifest itself in the absence of ideal-asdescriptive-model concepts that would provide the necessary macro and micro-mapping of that oppression, and that are requisite for understanding its reproductive dynamic.8 In political philosophy, defenders of “ideal theory” in Mills’s sense can maintain that by “fundamental” they mean something other than “important.” Explaining oppression and social change is important, maybe (ideal theorists could grant) equally important to questions that are “fundamental.” But by “fundamental,” the ideal-theoretic political philosopher means something else, perhaps “prior.” A case can be made that the questions of ideal theory are prior, because we need to know the principles of justice before we can even state the goal of social change. But to address the question of what the principles of justice are, one does not need to first address the question of social change. This, to the ideal theorist in political philosophy, grants priority to their inquiry.
Similar questions arise in the theory of meaning. Presumably we can envisage a world that is purely prosocial, in which there is only straight talk and incidental hustle, but no intentional hustle (a world lamentably bereft of stand-up comedy). But can one be intentionally hustled if there is no straight talk? Suppose that intentional hustle modally depends on the existence of straight talk, but not vice versa. Is that a basis of a case for restricting our theorizing in one mode to straight talk? Such priority need not exclude an investigation into the question of why we so often get hustled.
There are differences in the initial plausibility of objections and responses to idealizations in political philosophy and the philosophy of language. One (at least apparent) disanalogy is that there has never in human history been a well-ordered society. In other words, there has never in human history been a society that fits the ideal model in Rawlsian political philosophy. This is a broad problem for ideal models in political philosophy. But at least superficially, this is not as widespread a problem for idealizations about speech (though it is in the case of some of these idealizations, such as that of neutrality, which we argued was not so much impossible to achieve as it is incoherent). If so, it suggests that Mills’s concerns are more pressing in political philosophy than in the theory of meaning, as the idealizations are more extreme.9 But another disanalogy between the political philosophy case of ideal theory and the theory of meaning case suggests the opposing moral, that the methodological problem posed by the idealizations in the theory of meaning is worse. If we grant that even theorizing about social change presupposes a conception of justice, but not vice versa, then there is at least some understanding of priority that attaches to ideal theory. In contrast, it is no part of, for example, Grice’s project to provide an account of cooperative communication that then yields the concepts most central to the study of noncooperative communication.
From this perspective, the methodological problem in the theory of meaning is worse.
In both the case of political philosophy and the theory of meaning, there are legitimate concerns about the ideological function of ideal theory.
Mills’s contention in “ ‘Ideal Theory’ as Ideology” is that the ideal-theory approach to political philosophy has an ideological function, to conceal the importance of notions such as ideology, social change, and oppression. If ideal theory is “prior,” then these other notions are “secondary.” A similar concern arises in the theory of meaning. If what is “important,” “central,” and “prior” are the prosocial aspects of language pertaining to the conveying of information, then other effects of speech are “secondary.” This can obscure the ways in which aspects related to “tone,” expressive meaning, and connection to ideology are central to understanding the ways in which speech functions. If we only concentrate on the information conveyed by someone’s speech, we ignore the myriad other effects, both intentional and unintentional, in which speech impacts audiences. These concerns about the ideological effects of ideal-theory approaches are analogous.
Yet the disciplinary sociology of political philosophy and philosophy of language differ significantly.
10 Though to some extent the topic of oppressive or subordinating speech has not been accorded the same level of theoretical centrality as other work, it is hard to attribute the difference between the large amount of work on apolitical examples in the theory of meaning and the lesser amount of work on political examples to a problematic hegemonic ideological formation in the theory of meaning.
There is simply no dominant philosophy of language tradition akin to John Rawls’s program in political philosophy, or Jürgen Habermas’s, both of which share analogous ideal-theory assumptions. Contemporary theorists of meaning in philosophy and linguistics belong to many different traditions.
Our background for the most part is in the logic and semantics tradition.
In this tradition, natural languages are treated with the tools developed for the interpretation of the formal languages of logic and mathematics, which lack many of the properties of natural languages. Arguably, such formal languages lack what Frege termed “tone,” at least in the sense that they are not intended to model anything beyond truth and logical consequence.
Insofar as terms in these formal languages belong to speech practices, they are such marginal speech practices as to cast little general light on the study of human communication more broadly. It is the tradition of “ideallanguage philosophy.” By design, and despite important methodological changes in philosophy associated with the development of “ordinary language” philosophy in the mid-twentieth century, this tradition seeks to hide much of what happens when ordinary people talk.
Challenging the methodology of ideal political philosophy became central in political theory and political philosophy as academic disciplines in the 1990s and thereafter. But challenging ideal-language philosophy has a much older tradition. J. L. Austin’s 1962 How to Do Things with Words specifically targets idealizations about language. The central target of the book is “the descriptive fallacy,” that “the business of a statement can only be to describe some fact.”11 Austin argues ultimately that there is not even a distinguishable species of speech that is pure description, a claim that our model bears out. Ordinary-language philosophy is defined against “ideallanguage philosophy,” precisely in that the idealizations of the latter, the ideal properties of formal languages, were argued to misrepresent linguistic reality.
The affinities between Austin’s project and ours run deep. Our focus on practice rather than proposition is an extrapolation of his program in which conventions of language concern not merely representational techniques, but ways of doing things. Further, the concept of presupposition, the focus of part II of this book, is utterly central to Austin’s critique. Austin initially attempts to draw a distinction between performatives and constatives, where performatives have preconditions, and constatives just state something to be true or false.
But, as Austin eventually concludes, we can think of preconditions for performatives as their presuppositions, which constatives have as well.12 A practice has preconditions, and a description of the world presupposes practices that give that description its stability and meaning. We have argued, following Austin, that the same concept of presupposition is at issue in both cases. With Austin, we suspect that there are problems with the very division of speech into doing and describing, and our model bears this suspicion out. Like Austin, we recognize the need to operate with a concept of presupposition that does not depend on such a division.
9.2. The Content-Delivery Model In this section, we precisify the idealizations of “ideal-language philosophy” by focusing on something like Austin’s descriptive fallacy. In this book, we have caricatured the ideal-language philosophy approach as assuming models of communication involving a particular variant of Reddy’s conduit metaphor. Communication is seen as a cooperative enterprise of exchanging contents to interlocutors in the purpose of a common goal, and delivering them to each other in some sense inside little packets of words. Speech, according to the conduit model, is a contentdelivery system. If we think of speech as a content-delivery system, it is natural to develop tools, concepts, and resources that isolate the content-delivery aspect of language as precisely as possible. Aspects irrelevant to the content-delivery model can be filtered by idealizations. The use of language to evoke emotion or social bonding, or to negotiate social station and rank, distracts from the function of language to deliver information, and hence are aspects that should be filtered out of the core evidence for building a model of speech as a content-delivery system. Concepts, tools, and resources that are abstracted from speech as a content-delivery system can perhaps later be applied to these other functions of language, or they can be studied as peripheral uses of language.
How might one form idealizations that isolate speech as a contentdelivery system? It’s natural to focus on one speaker and one hearer—the person who delivers the content and the recipient of that content, and to think of them as engaged in a cooperative rational enterprise. It’s natural to abstract away from obstacles to smooth content delivery by speech—to think of the agents in our model as maximally open to the reception of content, with their intentions fully transparent, and to think of the various determinants of content as maximally mutually accessible. If the function of speech is to deliver information about a shared objective reality, it’s natural to think of conventions of language connecting words to neutral bits of information. If speech is content delivery, we can and perhaps should abstract from social roles and power relations. The conception of speech as centrally and most basically a conduit of information leads to what we can think of as the standard model, determined roughly by the following idealizations: DIALOGUE: A talk exchange is between one speaker and one hearer.
COOPERATIVITY: Speaker and hearer are cooperating in the service of a set of common interests.
RATIONALITY: Interlocutors are perfectly rational: they are computationally unlimited, reason scientifically and logically rather than emotionally, and have consistent preferences.
TRANSPARENCY: Utterance meaning, including presupposition and implicature, is characterized by a unique set of communicative intentions that are mutually and readily consciously recognizable. SHARED CONTEXT: Features of context relevant to interpretation must be mutually known in order that a unique content can be identified.
NEUTRALITY: (i) Conventions associated with words assign them a core of neutral and aperspectival meaning; (ii) at least some expressions are completely neutral, in the sense that perspective and attunement to social location are irrelevant to their meaning; and (iii) the neutral core of the meaning of a nonneutral expression is paraphrasable in neutral terms.
SOCIAL HOMOGENEITY: The linguistic community is socially homogeneous, and utterance meaning is computed without reference to social roles, affiliations, power relations, or personalities.
LANGUAGE HOMOGENEITY: Conventional meanings are determined primarily at a level of recognized languages, which may have millions of speakers. Speech practices of individuals or subgroups, registers, styles, differences from one communicative medium to another, and rhetorical frames of particular conversations, are not central.
PROPOSITIONALITY: Content is packaged into neat units, one proposition per utterance, and the primary point of communication is to convey these propositions with assertive speech acts.
FORCE: The primary level for studying communication is the illocutionary force of the utterance, which is a function of the underlying content. EXTENT: The individual utterance is the bearer of significant semantic properties. Properties of larger discourses, or temporally discontinuous exchanges, need to be considered only by extrapolation from the single utterance case.
These ideal-language idealizations result in a picture of speech designed to explain how it can be true or false. But we began this book by promising to explain how speech could also be harmful, violent, or dangerous. In crafting our theory, we considered as evidence a wide range of speech that is clearly nonideal according to this model. As we will try to show in chapter 10, although perhaps this is not in itself surprising, applying the tools and resources we derived from study of speech that is nonideal makes it much easier to explain what it is about speech that makes it harmful, violent, or dangerous. We contend that it is the ideal-language model itself that makes it mysterious how speech can have these properties, as well as many more positive properties, like being friendly, soothing, or clear. One might have thought that the last of these, the property of clarity, would be exactly what ideal-language models would be good for. Yet even here they fail, for as any educator knows, what is clear to one person is not clear to another, and this can only be understood once one drops idealizations like rationality, shared context, and social and language homogeneity.
While we have documented many cases in this volume of prior scholars who have denied particular idealizations, from Wittgenstein and Austin on, we have certainly gone further than others who, like us, work broadly in a tradition of analytic philosophers like Frege, Austin, Grice, and Montague.
Though there are surely ways in which we ourselves have simplified and abstracted from the complexities of real-world speech, we have nonetheless proposed a framework that eschews all eleven of the above idealizations.
Recapping material from the last eight chapters, let us rehearse some of the principal points that separate our proposals from much prior work.
DIALOGUE: Neither a crowd chanting in unison nor a dog-whistling politician intending to have different effects on different audiences fits the standard model. We have taken such data as central and discussed mechanisms for modeling communication at the level of groups. We make no assumptions about the assemblage involved in interactional events, or about the audience.
COOPERATIVITY: Our model assumes communication to be coordinative, but not inherently cooperative. Our practice-based model supports analysis of intended hustle, which we have argued to be a common feature of noncooperative communication.
RATIONALITY: We have argued that nondeliberative processes are central to communication, and we have suggested that the descriptive resonances of words (the way in which occurrences of words relate to features of the world) should not be privileged over emotional resonances.
TRANSPARENCY: We have argued against an idealization of straight talk, and we have suggested that hustle is central to communication. The practice-based model of communication does not make strong assumptions as regards intention recognition.
Recognition of intentions is an important aspect of communication, but it should not be assumed in the modeling of such central communicative processes as mimicry and emotional response.
SHARED CONTEXT: Our model applies in cases such as dog whistling where context crucially differs between audiences. We have also emphasized that much accommodation is gradual, and that accommodation can be divergent, with the net effect that mismatches between contextual assumptions are an intrinsic part of communication.
NEUTRALITY: We have argued that the concept of a neutral core is incoherent. All words come with perspective, and there is no such thing as an inherently neutral act.
SOCIAL HOMOGENEITY: Social resonances are as important in our model as descriptive resonances, and heterogeneity of groups and practices is at the core of our analysis of intergroup interaction, for example, processes of othering and the formation of echo chambers.
LANGUAGE HOMOGENEITY: We take as central situations of discursive conflict in which speech practices themselves are at stake, and in which groups compete for dominance of their preferred rhetorical frames. The language practices themselves are subject to both convergent and divergent accommodation, rather than being assumed constant across the community.
PROPOSITIONALITY: The haze of resonance associated with words is neither restricted to a single proposition, nor restricted to being propositional at all. Neither emotional resonances nor behavioral resonances nor resonances involving shift in attention are naturally thought of as propositions being exchanged between interlocutors.
FORCE: We have emphasized effects of utterances that go beyond the illocutionary force of the utterance, and indeed go beyond what is sometimes described as perlocutionary force (intended side-effects of an utterance). Every word in an utterance carries its own resonances, and the most important effects may be connected with accommodation of background rather than uptake of illocutionary force. EXTENT: Resonances of a practice are defined in terms of its extension, a history that dynamically changes with each utterance.
Each utterance helps shape the resonances of future utterances whether temporally close or distant. Oppressive speech consists of oppressive practices that cannot be fully understood by consideration of the effects of isolated oppressive speech acts alone. Many processes of accommodation and ideological change are gradual.
We hope it is clear that the framework we have set out does not rest on these particular idealizations. We accept that the exact idealizations of particular proposals in the literature vary widely. We also accept that the exact status of the idealizations also varies. Sometimes they are implicit in the choice of data, but never made explicit; sometimes, as with rationality, they are explicit simplifications but are not taken to be realistic commitments; sometimes, as with cooperativity in Grice’s work, they are regarded as communicative assumptions that structure reasoning whether they hold or not; and sometimes, as with neutrality in much post-Fregean work, they are strong theoretical commitments. At the very least, we take any of the idealizations we have listed to potentially restrict the empirical domain of inquiry in developing a theory of meaning in communication.
Thus, they are no innocent assumptions.
One way to object to a putative ideal model, determined by a set of idealizations, is by rejecting one or more of the ideals. But this does not mean that an ideal-theoretic methodology is inescapable. After all, it may be that each and every idealization, each way of limiting the data set for the study of speech, results in some unacceptable exclusions, ones that distort the theoretical concepts in a field built on speech situations that abstract from them. This is a possibility in the theory of meaning.
9.3. The Ideal/Nonideal Debate in the Theory of Meaning The validity of ideal theory as a method in philosophy is an issue that is essentially as old as the discipline of philosophy itself. In The Republic, Plato develops his theory of justice by describing what he regards as an ideal city-state, which, on his view, perfectly embodies justice. And toward the end of Book V, Socrates is asked to address how the ideal model he sketched could even be possible: “But, Socrates, if we allow you to go on like this, I fear you will never come to the point of discussing the matters you put aside in order to say all that you have just said.
Those matters, you recall, raised the question whether a city such as you have described could ever be a real possibility and how that possibility might be realized.
Were your city to be duly constituted, I am among those ready to admit that it would be a boon to its citizens.… Let us take it for granted that these and countless other advantages will accrue to the city so constituted. But let us leave off belaboring these points.
Excluding all other considerations, what should be done? The reverse.”13 The impossibility of realizing ideals is, we can say with some justification, the original nonideal concern for ideal theory. It is originally raised in the domain of political philosophy. And though Socrates takes this concern seriously, he has an answer that plausibly undermines it. Socrates argues that the concern can be dismissed, because it misconstrues the methodology of ideal theory in political philosophy, at least as he employs it. His goal in describing an ideal city is to allow a measure of how close we are to that (perhaps impossible) ideal of a just city. According to this defense of the ideal-theoretic method in political philosophy, providing an ideal allows one to have a measure to use to assess how close a given actual state is to that ideal. And that goal doesn’t require that the model determined as fully ideal has to be possible to realize at all.
The debate between ideal and nonideal methodologies has roiled political philosophy in recent years. But ideal-theoretic methodologies are employed in a variety of disciplines in philosophy, and debates and concerns about them are broader than just political philosophy.
And as we have already discussed several times, and as one would expect, the debate is different across areas of philosophy with different domains. Socrates’s position that the goal of political philosophy is to provide a measure of closeness to his ideal city is a claim about what should guide practice in political philosophy. But a similar view about what should guide practice in the philosophy of language would be manifestly absurd. The goal of philosophy of language does not even tangentially include helping to measure how close actual given discussions in language are to speech in an ideal model, whether it is an ideal model structured by the idealizations we discussed in the previous section, or one that is more general. Maybe this is what we are doing in political philosophy. But this just isn’t what we are or should be doing in philosophy of language.
For the sake of expediency, in the discussion to follow, which compares debates about ideal theory between disciplines, we will consider the idealizations that structure the content-delivery model as our example of an ideal model in the theory of meaning. However, our discussion is of broader significance, its conclusions relevant for ideal models that employ much less demanding idealizations.
On the ideal-theoretic approach underlying the content-delivery model, speech is, most centrally, a content-delivery system. If so, the thinking goes, it makes sense to form idealizations that allow us to focus on this function, abstracting from other functions. One kind of defense of ideal models is to argue that the idealizations are chosen carefully to limit attention to speech that is theoretically particularly generative. Because of the supposedly special nature of this subtype of speech, research just on it yields concepts and tools that are in fact applicable to functions of speech that are not obviously means of content delivery.
For example, one function of speech is to bond speaker and audience into a common social identity, which does not on the surface appear to be a function of speech that is explicable on the content-delivery model. But the advocate of the content-delivery model could defend its idealizations by maintaining that when speech functions to bond speaker and audience into a common social identity, it does so by means of sharing content—content about a common enemy, for example.
Bad speech, on this model, is a kind of misinformation.
Arguing that the idealized model provides surprising resources to account for speech that apparently falls out of its purview is not the only way to defend ideal models in philosophy of language, though it is perhaps the most difficult. Another kind of defense of ideal models can also be applied here, familiar even in the hard sciences. One could argue that the kind of speech that falls outside the ideal model is in some sense peripheral to theory, a kind of mere noise in its status as evidence for forming concepts. Yet a third defense of ideal models appeals to a distinction between the “pure” version of the inquiry, and an “applied” version.
For example, one could argue that the study of functions of speech outside content delivery is an applied study, to be contrasted with the pure study of the content-delivery function of speech, and as such may not be equally worthy of studying.
Some defenses of ideal-theoretic methodologies transfer to debates in other disciplines, or at least make some kind of minimal sense. Also, some concerns about ideal-theoretic methodologies transfer between disciplines.
Consider Charles Mills’s objections to two idealizations of Bayesian epistemology, which fall under Mills’s lens in “ ‘Ideal Theory’ as Ideology.” A first idealization critiqued by Mills is that of “idealized cognitive capacities”: The human agents as visualized in the theory will also often have completely unrealistic capacities attributed to them—unrealistic even for the privileged minority, let alone those subordinated in different ways, who would not have had an equal opportunity for their natural capacities to develop, and who would in fact typically be disabled in crucial respects.14 Second, Mills decries the assumption of “an idealized cognitive sphere”: Separate from, and in addition to, the idealization of human capacities, what could be termed an idealized cognitive sphere will also be presupposed. In other words, as a corollary of the general ignoring of oppression, the consequences of oppression for the social cognition of these agents, both the advantaged and the disadvantaged, will typically not be recognized, let alone theorized.
A general social transparency will be presumed, with cognitive obstacles minimized as limited to biases of self-interest or the intrinsic difficulties of understanding the world, and little or no attention paid to the distinctive role of hegemonic ideologies and group-specific experience in distorting our perceptions and conceptions of the social order.
15 Mills’s concerns about idealizations in epistemology carry over to the case of the idealizations of the standard model of communication. Speech can be oppressive via its content—if you claim that a group is essentially criminal, the oppressive effects can (largely?
entirely?) be captured by the content of your claim. But insofar as there are oppressive properties of speech that are not reducible to content, the idealizations function to mask them. If, for example, a practice of speaking is oppressive, that fact will be filtered out by the idealizations; a way of mocking the LGBT community by speaking in a certain pitch is oppressive, but not via its content. Manipulative speech can reinforce hierarchies, operating via nontransparent intentions that allow for plausible deniability (think of the use of dog whistles in political speech). And occluding manipulative speech by theoretical idealizations may be regarded as itself a kind of manipulation.16 Occluding manipulative speech is politically problematic, reinforcing existing social hierarchies by masking the existence of practices that support and maintain them. Mills’s concerns about two central highly idealized concepts in epistemology generalize to related concerns about some widely (though not universally) assumed idealizations in philosophy of language.
We have seen that versions of the impossibility argument against idealizations exist across various philosophical disciplines. But even when the same basic objection to idealizations does transfer across disciplines, targeting other idealizations with the same problem, the replies may not. As we have seen in Socrates’s defense of the “impossibility worry” for ideal models in political philosophy, it is possible to respond to this objection by appealing to the aim of political philosophy. However, nothing like his response has any plausibility in the philosophy of language.
The “impossibility objection” arises in epistemology as well, and here too the most plausible replies are nonstarters in the theory of meaning. Here is a difficult version of the impossibility objection, arising against idealizations in Bayesian epistemology, as stated by David Christensen: As many people have pointed out, attaining probabilistic coherence is far beyond the capacity of any real human being. Probabilistic coherence, after all, requires having full credence in all logical truths —including complicated theorems that no human has been able to prove. It also places constraints on beliefs about logically contingent matters—constraints that go beyond human capacities to obey.… The fact that this sort of “logical omniscience” is built into probabilistic coherence has led many to doubt that coherence can provide any sort of interesting normative constraint on rationality.
17 There are of course also, as Christensen proceeds to point out, similar idealizations with deductive rationality; for the same reasons, we do not know all the deductive consequences of our beliefs.18 In the final paragraph of Putting Logic in Its Place, Christiansen replies to the version of the impossibility objection as it arises against idealizations in Bayesian epistemology: Furthermore, philosophy in general, and epistemology in particular, need not be directed toward external practical ends. We surely may philosophize because we hope (perhaps optimistically) to help people improve themselves cognitively. But just as surely, epistemologists need not restrict their efforts to improving our educational system, or to producing popular manuals for cognitive self-help. We may philosophize because we want a better understanding of ourselves—of our cognitive natures and our situation in the world.
We may philosophize because we want a better understanding of rationality itself.19 We find it problematic to pose the options as being between an epistemologist who restricts their efforts to improving our educational system, an epistemologist who produces popular manuals for cognitive selfhelp, and an epistemologist who wants a better understanding of rationality itself. Christensen’s way of setting up the options imposes immediately a distinction between the applied study of a subject matter, and the pure study of that subject matter; only the third kind of epistemologist would count as a pure epistemologist. This distinction has proven to be politically problematic. In the case of all such disciplinary distinctions, the study considered to be pure is associated with more prestige.
Christensen’s defense of the unattainable ideals of Bayesian rationality is that it is an exercise in a kind of idealized Platonic inquiry, an attempt to grasp the form of rationality.
This project, he argues, requires abstracting from the limitations of reality that are obstacles to the manifestation of ideal rationality. However, as in the case of Socrates’s reply to the impossibility objection to his ideal model of a city in The Republic, the defense that Christensen envisages for ideal-theoretic epistemology just doesn’t get to the starting line in philosophy of language. The expressive functions of speech, the functions of language to bond and share identity, are genuine functions of speech—they are not in any sense “practical applications” of a pure theory of speech, for example as content delivery. Moreover, the theory we have developed makes this vivid. In that theory, cognitive attunement is at the same level as affective attunement, or attunement to identity. If the model of meaning we have developed is on the right track, then standard defenses of ideal theory in epistemology cannot possibly be replicated here. It’s hard to think of idealizations in the theorization of communication that could be defended on the grounds that they lay language bare by focusing on its true function. Rather, too many of the idealizations in the theory of meaning function to cover up important roles that language plays in society, leaving an impoverished and artificial theory of communication incapable of being extended beyond toy examples.
In recent work, Herman Cappelen and Josh Dever end up at a position that is in some ways similar to our own, and in some ways distinct: Our job for this volume was to clarify the distinction between Ideal and Non-Ideal Philosophy of Language. We’ve ended up rejecting the distinction. However, maybe there’s another, closely related distinction that is useful. This volume has in its title the phrase “social and political philosophy of language.” What does that category pick out and what is it contrasted with? At the risk of appearing to be incurable curmudgeons, we are going to also reject that category as fairly useless. What could it possibly pick out?
Maybe it’s an attempt to pick out speech by people who are politicians or people who talk about political/social topics. If that’s the remit of social-political philosophy of language, then it should be concerned with sentences like: “We should increase the sale tax on cigarettes because that would give the city more money for schools and it could reduce cancer rates.” “The price of parking on city streets should be increased because it is not fair that public property should be rented to car owners for very little money.” That’s what the vast majority of “political speech” is like. It’s just “ordinary speech.” It’s not special in any way. It doesn’t constitute a distinct subset of speech.20 Though our arguments differ, we agree with Cappelen and Dever that there is no sharp distinction between ideal and nonideal philosophy of language, and we have shown in detail that it cannot possibly be an entirely parallel debate to the ones in epistemology or political philosophy. We have argued that idealized philosophy of language cannot be argued to be prior to nonidealized philosophy of language, and indeed this is in line with the fact that work in the ordinary-language tradition has always actively questioned simplifying idealizations that see language as a purely representational system. But it is an idealization to assume that categories must be sharp in order to be useful, and the question of whether nonideal philosophy of language is sharply delineated from ideal philosophy of language is distinct from the question of whether the category “nonideal philosophy of language” is a useful one. We do not take a stand on this latter question here, but note that the editors of the handbook in which Cappelen and Dever’s argument appears, the Routledge Handbook of Social and Political Philosophy of Language, apparently found the distinction to be useful, at least as an organizational device.
In the opening of the Routledge Handbook, the editors, Rachel Sterken and Justin Khoo, define the remit of the volume as bringing together “work on how language shapes and is shaped by social and political factors.”21 We agree with Cappelen and Dever that social and political language is not a separate category of language. It is a central tenet of this volume that all language is social and political. But we disagree entirely with their conclusion that “social and political philosophy of language” is “fairly useless.” Rather, we presume we are in agreement with Khoo and Sterken.
“Work on how language shapes and is shaped by social and political factors,” which constitutes social and political philosophy of language, has become a lively subarea of academic research, building especially on the work of feminist and race scholars; we take it that progress in this subarea is essential to progress in philosophy of language.
We have argued that some of the most well-known defenses of ideal models in political philosophy and epistemology are simply absent in the case of the theory of meaning. There is no good case, for example, that the content-delivery model or the prosocial model of speech is “prior” to the study of other aspects of speech. And there is no good case that anything like the content-delivery model is “ideal,” and the use of speech in propaganda is a mere “practical application.” This doesn’t mean, of course, that the theory of meaning can do without idealizations. But it does perhaps help explain why nonideal approaches have a long and distinguished history in twentieth-century philosophy of language.
1. Wittgenstein, Tractatus Logico-Philosophicus, par. 81.
2. Material from this chapter first appeared in 2018, as Beaver and Stanley, “Toward a Non-Ideal Philosophy of Language.” 3. For critiques of neoclassical economics’ assumption of rationality, see, e.g., Herbert Simon’s argument for a notion of “bounded rationality” in “A Behavioral Model of Rational Choice”; Hollis and Nell, Rational Economic Man; Tversky and Kahneman, “The Framing of Decisions.” For but a few critiques of rationality assumptions in philosophy, psychology, law, and other social theory, see Cherniak, Minimal Rationality; Elster, “When Rationality Fails”; Thaler and Sunstein, Nudge.
4. Mills, “White Ignorance,” 14.
5. Anderson, “Feminist Epistemology and Philosophy of Science,” section 1.
6. Mills, “ ‘Ideal Theory’ as Ideology.” 7. Rawls, A Theory of Justice, 4.
8. Mills, “ ‘Ideal Theory’ as Ideology,” 168–69.
9. The concession that there are many actual talk exchanges that meet the standard model is just for the sake of argument. As will emerge below, we are not ourselves convinced of this.
10. Historians of philosophy have argued that similar ideological exclusions caused by disciplinary formations functioned to exclude work in modern philosophy by women, e.g., O’Neill, “Disappearing Ink.” 11. Austin, How to Do Things with Words, 1.
12. Austin’s brief discussion of the relationship between preconditions and presuppositions culminates as follows: “Here we might have used the ‘presuppose’ formula: we might say that the formula ‘I do’ presupposes lots of things: if these are not satisfied the formula is unhappy, void: it does not succeed in being a contract when the reference fails (or even when it is ambiguous) any more than the other succeeds in being a statement. Similarly the question of goodness or badness of advice does not arise if you are not in a position to advise me about that matter” (How to Do Things with Words, 51).
13. Plato, The Republic 188–89 (471c–e).
14. Mills, “ ‘Ideal Theory’ as Ideology,” 168.
15. Mills, “ ‘Ideal Theory’ as Ideology,” 169.
16. It is not just manipulative speech that these idealizations occlude—social bonding and the creation of in-groups and out-groups are processes that do not simply work by the delivery of content. And on the model we provided in parts I and II of this book, these functions of speech are not derivative from content delivery, or at least they are not derivative from a process that operates anything like that in the paradigmatic models of content delivery described by the idealizations we considered in the last section.
17. Christensen, Putting Logic in Its Place, 150–51. 18. This is a problem with the idealized model of Bayesian epistemology that (nonaccidentally) has a direct correlate in the case of the possible-worlds theory of meaning, as discussed in the context of fragmentation in section 3.2.
19. Christensen, Putting Logic in Its Place, 178. 20. Cappelen and Dever, “On the Uselessness of the Distinction between Ideal and Non-Ideal Theory,” 101. Their view, which perhaps is in part a riposte to an early presentation of our work on idealization at the 2017 conference on Philosophical Linguistics and Linguistical Philosophy (PhLiP), is echoed in the first chapter of their introductory textbook, Cappelen and Dever, Bad Language.
21. Khoo and Sterken, Handbook of Social and Political Philosophy of Language, i.
PART IV Oppression and Freedom
CHAPTER TEN Harmful Speech Official language smitheryed to sanction ignorance and preserve privilege is a suit of armor polished to shocking glitter, a husk from which the knight departed long ago. Yet there it is: dumb, predatory, sentimental. Exciting reverence in schoolchildren, providing shelter for despots, summoning false memories of stability, harmony among the public.… Oppressive language does more than represent violence; it is violence; does more than represent the limits of knowledge; it limits knowledge.
—TONI MORRISON1 Officialese [Amtssprache] is my only language.
—ADOLF EICHMANN, AS QUOTED BY HANNAH ARENDT2 WE BEGAN THIS BOOK by promising that the tools and concepts we develop would explain how speech can be harmful, as straightforwardly as the ideal language model explains how speech can be true or false.
In this chapter, we attempt to deliver on this promise.
In chapter 7, we discussed the ideal of neutrality. Whatever oppressive language is, it is not neutral.3 The trouble is that the ideal of neutrality has obscured the nature and indeed the ubiquity of oppressive language. Yet the centrality of the ideal of neutrality to the dominant ideology of academic philosophy of language and linguistic semantics, what we have called the standard model, has meant that most analysts base their work on the assumption that all language has a neutral core. That has led to a tendency to ignore much nonneutrality in language, pushing the study of nonneutrality to pragmatics, and to some extent exiling it so that many types of nonneutrality, effects of language other than for conveying factual information about the world, are only studied in separate academic disciplines altogether, such as communication studies and psychology.
We return below to the topic of neutrality, specifically how the ideal of neutrality is implicated in masking harm across three categories of speech we discuss in this chapter: slurs, genocidal speech, and bureaucratic speech.
Slurring is one paradigm of nonideal language. With a single word, a speaker can perform an act that is at once noncooperative, centered on an in-group/out-group distinction, indicative of differential power and status, and so emotionally charged that it may foreclose informationally fruitful discussion. To the extent that a slur provides information, what it says is not easily negotiable or debatable. And although slurring is, by its nature, typically a transparent, in-your-face act, it will turn out that a certain amount of hustle is present in all slurring, and that in the case of slurring using words that are not normally regarded as slurs, there is an additional element of hustle connected to the presence of plausible deniability.
We began this book with Cleon’s contribution to the Mytilenean debate, where Cleon argues that Athens should massacre the entire population of Mytilene for rebelling. Cleon’s speech, the paradigm of demagogic speech in the ancient world, is genocidal speech. We see again today, in Russia’s justification for its invasion of Ukraine, these very same tropes, attesting to their permanent power. Cleon explicitly represents democratic values as inimical to his goals. Genocidal speech is the original paradigm of antidemocratic speech, the central example of demagoguery in the ancient world.
Why, however, do we place bureaucratic language in the category of oppressive speech?
We begin by addressing this question—and the problematic role the ideal of neutrality characteristically plays in masking harm.
10.1. Oppressive Language and the Ideal of Neutrality In looking at slurs, genocidal language, and bureaucratic language, it is tempting to treat the first two as belonging together, perhaps lumped together under the label “hate speech,” and the third as belonging to a distant and unrelated category, similar only insofar as all three are political.
It is important to avoid this temptation. First, what is remarkable about much of the language of genocides is its unremarkability. That is, The Banality of Evil, to adapt from Arendt’s famous title, is reflected in the banality of much of the language of evil.4 As she says, None of the various “language rules,” carefully contrived to deceive and to camouflage, had a more decisive effect on the mentality of the killers than this first war decree of Hitler, in which the word for “murder” was replaced by the phrase “to grant a mercy death.” Eichmann, asked by the police examiner if the directive to avoid “unnecessary hardships” was not a bit ironic, in view of the fact that the destination of these people was certain death anyhow, did not even understand the question, so firmly was it still anchored in his mind that the unforgivable sin was not to kill people but to cause unnecessary pain.5 To put it another way, let us borrow instead the title of the historian Christopher Browning’s well-known study of the history and psychology of people who became part of the Nazi genocide, as members of a battalion of Ordnungspolizei (literally: order police) in occupied Poland. What could be more ordinary than maintaining order according to the rule of law? The problem was that the order they were empowered to maintain implied mass murder. Just as it is remarkable how in a genocide Ordinary Men can become twisted into performing the ugliest of atrocities, so too it is remarkable how in a genocide what we might think of as ordinary language can become twisted.6 Second, bureaucratic language, far from being an unrelated category, is in fact often crucial to genocidal regimes, and to discriminatory ideologies more generally.
We have already argued that language is never neutral. Let us go further: perhaps it is a mistake to ever think of language as ordinary. To say that language is ordinary is to make a claim about what is commonplace in a particular community of practice. But what seems ordinary to those in the midst of a genocide should seem extraordinary to us. So, we must ask ourselves whether the apparent ordinariness of a turn of phrase reflects inherent ordinariness, or whether it in fact reflects our own insensitivity to the role that language plays within ideologies that are so familiar to us as to appear unremarkable. In The Language of the Third Reich, Victor Klemperer states that the first three words that he noticed as being specific to Nazi vocabulary were Strafexpedition (punitive expedition), Staatsakt (state occasion), and historisch (historical).7 The first, used to describe attacks by groups of brownshirts in the early days of Nazi Germany, is intrinsically violent. This is seen in Klemperer’s discussion of what we would term the word’s resonances: “For me the word Strafexpedition was the embodiment of brutal arrogance and contempt for people who are in any way different, it sounded so colonial, you could see the encircled negro village, you could hear the cracking of the hippopotamus whip.”8 The imagery has visceral power, and perhaps conveys the sense in which the word was not merely violent, but inextricably linked to discriminatory ideology. However, the other two words seem innocuous.
Klemperer describes the propaganda minister, Goebbels, as having staged “an almost incalculably long series” of Staatsakten. To use this word was not merely to describe the type of occasion that inevitably happens in a state. It was to imbue it with national significance and to force attention toward it. As Klemperer makes clear, both the word Staatsakt and the occasions so described, frequently surrounding the coffin of a war hero, were advertising for the indomitable power of the Nazi state: “The splendour of the banners, parades, garlands, fanfares and choruses, the allembracing framework of speeches, these all remained constant features and were undoubtedly modelled on the example of Mussolini.”9 The application of the word historisch would appear at first to be up to historians, a group who might be imagined, or who might conceivably imagine themselves, as presenting both a dispassionate view of events that have occurred over large time scales, as well as the logic and significance of those events. Yet it is also a familiar idea that history is intrinsically ideological, so we should not be surprised that the application of the term historisch became an important tool of Nazi propaganda. It is the goal of the propagandist to make history. Here is what Klemperer says: Which brings us to the word that National Socialism used from beginning to end with inordinate profligacy. It takes itself so seriously, it is so convinced of the permanence of its institutions, or at least is so keen to persuade others of that permanence, that every trifle, however insignificant, and everything that it comes into contact with, has a historical significance. Every speech delivered by the Führer is historical [historisch], even if he says the same thing a hundred times over, every meeting the Führer has with the Duce is historical, even if it doesn’t make the slightest difference to the existing state of things; the victory of a German racing car is historical, as is the official opening of a new motorway, and every single road, and every single section of every single road, is officially inaugurated; every harvest festival is historical, every Party rally, every feast day of any kind; … [the Third Reich] views every single day of its life as historical.10 The word historisch can be presented as if it is an objective assessment, but it was simply a mask used by those who wished to draw attention to whatever suited their purpose as symbols of the power and success of the Nazi regime. Nazi military defeats were not historisch.
A recurring theme of Klemperer’s writing is the special language of Nazi bureaucracy, and in particular the use of mechanistic metaphors.11 As Klemperer describes it, the use of mechanistic language to describe humans emphasized the fact that they were expected to act without thinking.
Klemperer describes much of this metaphorical language as already present in abstract descriptions of institutions, but being used in a new way, to apply to individual people.
Consider the English word “alignment.” It has a purely physical meaning that would be relevant in mechanical and architectural tasks. But people can also be aligned, and they can be forced into line, literally or metaphorically: The explicit mechanization of the individual himself is left up to the LTI. Its most characteristic, and probably also earliest, creation in this field is “gleich-schalten [to force into line].” You can see and hear the button at work which forces people—not institutions and impersonal authorities—to adopt the same, uniform attitude and movements: teachers in various institutions, various groups of employees in the judiciary and tax authorities, members of the Stahlhelm and the SA, and so on, are brought into line almost ad infinitum.
12 The word gleichschalten is banal. But the banality of bureaucratic language is precisely what masks its insidious power. It is well established that under the Nazi regime, violence was bureaucratized. Klemperer shows us that the language of violence, too, was bureaucratized.
We now turn to how terms that might be seen as just the ordinary way of referring to a group, and hence supposedly neutral, can mask the way these terms are used as slurs. We approach this indirectly, by first considering how the concept of neutrality has been applied in the case of slurs. Building on recent work in the area, we will suggest that the mistaken assumption that slurs have neutral counterparts is intimately linked to a mistaken view of group-denoting terms more generally. We will offer an analysis of groupdenoting terms that depends on the presence of multiple communities of practice in contact with each other, but with differing ideologies. With a toxic enough background ideology in the context, any group-denoting term can be transformed into a tool for slurring.
We argued in section 7.3 that it was incoherent to treat evaluative words as having neutral counterparts, or for that matter, a neutral core. So it is with slurs. Like other evaluative words (say, “generous”), they lack a neutral core. As for what the neutral counterpart of an act of slurring would be, that is an odd thing to ask for. If you neutralize an insult, the one thing you do not have is an insult. Yet the vast bulk of the prior literature on slurs assumes that slur words have neutral counterparts.
Scholars who take slurs to have a neutral counterpart also commonly assume that the meaning of that counterpart constitutes the neutral core meaning of the slur. Such scholars generally postulate some special add-on to the core, which is variously described as a conventional implicature, a presupposition, or tone. These analyses miss the mark. We agree, rather, with three scholars who have (independently) argued against the neutralcounterpart view: Lauren Ashwell, Heather Burnett, and Jennifer Foster.
13 As we argued in chapter 7, evaluative words like “stupid” or “ugly” cannot possibly have neutral cores. It is implausible that their meaning consists of a sense that lacks evaluativity and a completely independent tone that is evaluative. We see the general argument against evaluative words having neutral cores as applying equally to slurs, as well as to emotively charged expletives. We don’t think linguistic theory would be advanced by a claim that an expletive like “Fucking hell!” has a non-emotive, purportedly neutral core, perhaps something like “I am taken aback!” and a patina of tone. In the case of evaluative words, we argued that if there is to be any core, or sense, then it should correspond to a concept that is intrinsically evaluative, that is, an evaluative category. Let us go further in the case of slurs. Slurs do not just have bad tone, or bad presuppositions, or bad conventional implicatures. Their core meaning is a bad category: the category picked out by a slur is inherently negative within the ideology that gives that slur meaning.
We must be careful in naming slurs, but to illustrate, let us consider “fat cat” and “commie.”14 Both have uses in acts of derogatory name-calling, as well as within in-group discussion to slur others who are not present. Yet, we take it that at the time of writing neither of them carry such a kick that its mere mention would be deeply upsetting to many people. The phrase “fat cat” exists within an ideology that views some aspects of the prevailing economic system, presumably including wealth inequality, as problematic.
The phrase then has as its extension some set of individuals with an inappropriate amount of wealth or power. To predicate the term of one of these individuals is to pass a negative judgment on them for this reason. It is pointless to look for a neutral version of “fat cat.” One might claim, for example, that “person” or “person with any amount of wealth and power” was the neutral core, but this would just beg the question of how predication of the term “fat cat” could come to reliably pick out a narrow segment of the population. The reference is clearly to a set of individuals not just with any amount of wealth and power, but to a set of individuals with excessive wealth and power, and the word “excessive” is normative.
Likewise, to call someone a “commie” is to judge them negatively, to judge them as belonging to a category that is despised within the ideology of the person who passes judgment. Now here one might object that surely “communist” is a neutral counterpart. We disagree, for the simple reason that there is nothing in the least neutral about the term “communist.” And in particular, within the communities of practice that use “commie” as a slur, it is clearly the case that both communism and communists are seen as inherently problematic. To be a “commie” or a “communist,” according to these ideologies, is just as bad as being a “Marxist,” and not much different from being a “looney Marxist” (cf. the Dan Patrick quote in section 3.5).
None of these terms neutrally pick out a set of individuals who subscribe to a particular set of political ideals or live within a national system characterized by such ideals.
It can be argued that all of the above terms related to communism are not merely nonneutral, but in fact have racist resonances, and are regularly deployed as a form of racist dog whistling, as we will discuss in the next section. In the ideology within which “commie” is commonly used, it not only categorizes someone negatively, but characterizes them as belonging to a category that is partly defined in terms of race and racial sympathies.
Hitler, McCarthy, and many others have tied communism to Jews, and Jews were disproportionately represented as targets of McCarthy’s Red Scare.
But the racial overtones of communism-related words are more complex.
The line “Some say it’s a communist plot” from Nina Simone’s powerful protest song “Mississippi Goddam” concerns a separate but interwoven racist thread connecting African-Americans, and the goals and successes of the Civil Rights movement, to communist agitation. Martin Luther King Jr.
was regularly accused of being a communist. All of this makes sense once one considers the one thing that all these groups, that is, Jews, Blacks, and communists, have in common from the point of view of extreme right-wing discriminatory ideologies: they are all out-groups. Lumping out-groups together into one vast plot is at the heart of conspiracy theorizing. It makes the story about the enemy infinitely simpler than the confounding detail of reality.
There is nothing neutral about the word “communist,” or about the concept it picks out, within the discriminatory ideologies that have also offered us “pinko,” “lefty,” and sneering uses of “comrade.” Similar comments apply to the supposedly neutral counterparts of explicitly racial slurs, like the N-word. Neither “Black” nor “African-American” nor any other phrase is a neutral counterpart. The N-word is deployed within ideologies in which none of these words are neutral. In fact, although within the practice the three different terms (the N-word, “Black,” and “AfricanAmerican”) are involved in differing categories, all of them may have roughly the same denotation within a particular discriminatory ideology.
Within such an ideology, all three terms could pick out a despised out-group category whose extension is understood in essentialist terms, whose members are associated with various stereotypical features and are treated in various oppressive ways. To reprise our comments in chapter 7, while we are not sure that the idea of a neutral concept is even coherent, we certainly doubt that any racial or ethnic categorization could possibly be neutral, since perspective and social location are always relevant to the functioning of communicative acts that involve race and ethnicity.
One of Foster’s points is that the supposedly neutral counterparts to slurs can in fact be used to slur. That is, people sometimes use “Black,” “Mexican,” and “Jew” in an insulting way. Since we don’t think they are neutral, let’s just call common names for racial, ethnic, and religious groups that are often used without derogation such as these standard socioethnic group terms. What, then, is the relationship between standard socioethnic group terms, which are not normally thought of as slurs, and the corresponding terms that are standardly recognized as slurs? We suggest that the difference is that whereas the slurring terms are actively used only within communities of practice with discriminatory ideologies, the parallel standard socioethnic group terms are used within broader communities of practice for which the ideologies are not discriminatory to the same extent, if at all.
As a result of its parallel use within a practice in a salient nondiscriminatory ideology, the use of a standard socioethnic group term does not in and of itself provide evidence as to whether it is being used within a discriminatory ideology. Nonetheless, it will sometimes become apparent that this is the case. It might become apparent that the speaker is using the term as a slur, perhaps to name the very same role in an ideology that they might otherwise have used the slur term for. That could be because the speaker has signaled their racist ideology through separate means, whether the speaker has adopted a sneering tone or is making an inherently racist claim, or whether the speaker has chosen to refer to a socioethnic group at a point in a conversation where membership of that group would not be relevant outside of a racist ideology. It is, indeed, inevitable that sometimes standard socioethnic group terms will be used to pick out despised roles in discriminatory ideologies, because use of these words allows racists to slur while still maintaining some degree of plausible deniability.
For example, there are antisemitic communities of practice within which slurs are used to refer to Jewish people, and within these communities of practice the word “Jew” might be used coextensively, to label roughly the same people who would be labeled using the slur terms, or might be used to label a slightly different superset of people as would be labeled using the slur term. In either case, the use of “Jew” by people in this community is disdainful, and the disdain may often be manifest among interlocutors. Such an antisemitic community of practice can exist within a much larger broader community of practice, say speakers of American English. Let us assume that the broader community of practice is not, collectively, antisemitic.
Then uses of the word “Jew” within this broader community of practice will usually not be acts of slurring, and will not reveal or signal disdain. It is only when a use of “Jew” is a performance of the practices of the antisemitic subcommunity that the use can be a slurring act. The question of when exactly we should take a use of “Jew” to represent the antisemitic practice, and when we should take it to exemplify the practice of the broader community, which by assumption is not intrinsically antisemitic, is always going to be difficult. That is where plausible deniability comes in.
The difficulty in establishing which practice a use of the word “Jew” tokens, and hence what resonances are associated with that use, is one reason why an antisemite might on occasion choose to slur people using the term “Jew” rather than using a word that is specialized only for slurring.
The use of standard socioethnic terms by racists can be paralleled to the use of dog whistles. Both depend on the existence of two communities of practice with differing ideologies. A dog whistle has complex resonances, but different audience members are sensitive to different resonances and respond to them in different ways, which allows dog whistling to signal one thing to members of one community, and something else to members of the other community, and yet allow the speaker to maintain plausible deniability. A slurring use of a standard socioethnic term might similarly be described as having resonances that audience members are sensitive to in different ways. It follows that it would be misleading to describe standard socioethnic terms or dog whistles as “ambiguous” between different meanings, for this would suggest that on any occasion of use, one meaning was the correct and intended meaning, and the other was incorrect and not intended.
The true pragmatics of the discursive situation is still more complex. For as with dog whistles, the existence of the two ideologies is often well known to people on both sides of the ideological divide. In our terms, there are strong second-order attunements to both ideologies among at least some members of each community. When the racist slurs using a standard socioethnic term, they may intend a double entendre, or indeed succeed in producing it with no such intention. Some see their Janus-faced character, while others fail to see it, or avert their eyes. The metaphor of Janus is apt, for Janus was the Roman god of duality and portals between worlds. It might then be said that disguised slurring practices and dog whistles hang over slightly hidden portals between ideological communities, in which case canonical slurs are highly visible gates to ideological underworlds, opened to some and shut hard in the face of others.
There can be clear strategic value in using the terms “welfare” or “law and order” in ways that speakers know will be recognized as racist by many (if not all) in the less racist community of practice; the same is true when a politician uses a standard socioeconomic term in a racist way. Such uses draw joint attention to inherently divisive issues. These uses will be used strategically when someone in a racist community of practice sees utility in divisiveness. As we saw in chapter 6, such divisiveness is often not just incidental to political movements based around discriminatory ideologies, but is a core strategy. It is a way to form an antagonistic social-identity group, where membership is defined by opposition to another group. As the American far-right strategist Steve Bannon said in 2018, “Let them call you racists, let them call you xenophobes. Let them call you nativists. Wear it as a badge of honor.”15 The relationship between standard socioethnic group terms and their distinctively slurring counterparts is a complex one. Within the same ideology, they may possibly have very similar meanings, but across ideologies they have at best the status of being something similar to translation equivalents. Since the ideologies are distinct (one being significantly more discriminatory than the other), Quine’s problem of radical translation applies. Though certainly a term in one language might be a close counterpart of a term in another language, indicating that the terms play similar roles in the ideologies of their respective communities of practice, something is always, as they say, lost in translation.
The idea of untranslatable words is a familiar one, although it is a curious fact that online lists of untranslatable Yiddish words are typically paired with definitions. In this strange hinterland of borrowed meaning, we note a similarity between borrowings and slurring uses of standard socioethnic group terms. Many Yiddish words are borrowed into English, and thence become enmeshed in entirely new communicative practices, so that it is doubtful that they mean the same thing as they did in their original language setting. It is curious that something may be lost when a word is “translated” into itself. How the resonances of a Yiddish borrowing in English are related to the resonances of the original depends on how the history of the practice of using the word is sliced and diced. And there is no fact of the matter. There are just different language users with their own partial awarenesses of different parts of the extension of a practice that has crossed between worlds.
So it is with slurring uses of standard socioethnic group terms, so that racists and nonracists use the same standardized words for races or ethnic groups. Perhaps one should rather describe this situation as being one where it is not well defined whether the word used by racists and nonracists is the same; what is clear is that the usages belong to distinct subpractices in different subcommunities, and that the resonances of these subpractices are distinct. Likewise, left-wing Americans and right-wing Americans have rather different understandings of words like “communist,” “liberal,” and “fascist,” both in the sense that they would understand the terms as applying to different sets of people, and in the sense that members of one group might use them as slurs when the other would not. Different people have different awarenesses of the resonances the words carry. It is a discomforting situation, and that is part of the reason they are so used.
Let us mention one further example that is revealing of the intricacies of words that are not exactly slurs as the term is usually understood: the word “boy.” At first blush, at least if we were not writing in a sociopolitical context, the word “boy” might seem as neutral, ordinary, and banal as the word “dog.” Yet like “dog,” the word “boy” has distinctive resonances. Any use of a diminutive form or reference to someone as if they were a child implies a lack of respect for them. And within certain discriminatory ideologies, this particular word is involved in a very special practice indeed, one interwoven with the awful history of slavery and oppression in the United States. Neither is this practice dead. President Donald Trump’s first attorney general, it turns out, repeatedly addressed at least one black assistant US attorney as “boy.”16 To the extent that any word is innocent, in the sense of lacking specifically discriminatory resonance, it may easily be sullied in the mouth of the racist.
10.2. Slurring The resonances of swearwords and slurs (particularly when used in acts of slurring) impinge themselves on the hearer like an inebriated gatecrasher at a party, drawing attention to themselves, brash, enervating, obdurate, and even violent. Slurs have the power to wound, perhaps in a way that demands legal remedies.17 One slur in particular has been described as “the nuclear bomb of racial epithets.”18 In reaction to the use of this slur by White TV personality Bill Maher, the rapper Ice Cube, who is Black, compared the slur’s power to the stab of the knife, observing that it wounds independently of the intentions of the speaker: I think there’s a lot of guys out there who cross the line because they’re a little too familiar, or they think they’re too familiar. Or, guys that, you know, might have a black girlfriend or two that made them Kool-Aid every now and then, and then they think they can cross the line.
And they can’t. You know, it’s a word that has been used against us. It’s like a knife, man.
You can use it as a weapon or you can use it as a tool. It’s when you use it as a weapon against us, by white people, and we’re not going to let that happened again … because it’s not cool.… That’s our word, and you can’t have it back.… It’s not cool because when I hear my homie say it, it don’t feel like venom. When I hear a white person say it, it feel like that knife stabbing you, even if they don’t mean to.19 Philosophy and law scholar Patricia J.
Williams has said of that same slur: “It hits in the gut, catches the eye, knots the stomach, jerks the knee, grabs the arm.”20 Such effects seem almost magical. How do we explain them in terms of conventional properties of words?
Slurs have been at the center of discussions of harmful speech because of their distinctive power to harm. But slurs have attracted such attention to theorists in philosophy of language and linguistics because they seem to have the surprising property of being able to do such harm as words. Given what we have called the standard model, it is not clear how this could be so. The most standard route taken in literature on the meaning of slurs, discussed below, emphasizes the role of the hearer. On this type of view, even someone who utters a slur to the face of a victim in a direct act of name-calling at most expresses or conveys their own negative attitude, thus derogating the victim. That is where the conventional meaning of the slur stops. Any further negative effects on hearers would have to result from pragmatically mediated processes that lead them to take offense.
While we don’t doubt that pragmatic reasoning comes into play, as with all human communication, such an account puts the jackboot on the wrong foot, and dangerously so.
Although we don’t for a moment imagine that this has been the intent of past scholars, it seems to us that this way of looking at the harms of hate speech has much the same abstract structure as victimblaming rhetoric does in general. Our argument will be that slurs are not merely offensive but injurious.
21 We take slurs to be imbued with dangerous power, and yet we do not think them magical.
Quite the opposite. As we will suggest below, slurs are about as ordinary as language gets.
The properties of slurs, such as their powers to wound, to grab attention, and to reveal ideology, can be analyzed in terms of resonance, but that is not in itself remarkable: it is how we propose to analyze all conventionalized linguistic meaning in this volume.
If we are on the right track in proposing our resonance-based framework, then no special grammatical mechanisms are needed for slurs. The job of this section of the chapter is to convince the reader that insofar as slurs are a sound test case we are indeed on the right track.
Before proceeding, let us briefly recap the features of our framework.
Words have resonances, all the things that are found in the contexts in which they are used.
People and collections of people have attunements, which are ways of behaving, thinking, and feeling about things. When people have attunements to a word, that means not only that they have theoretical knowledge of what the word means and how it fits into the grammar, but also that they have practical knowledge. That means both that they use the word and react to use of the word in ways that are appropriate to the community of practice within which the word is found.
As we have set things up, resonance and attunement are two sides of the same coin. To be attuned to any practice is to have a tendency for change of behavior and state in accord with the resonances of the practice. Our notion of harmony concerns the fact that people feel dissonance when they sense a clash of salient attunements, and consonance when they feel alignment and coherence. An ideology is then a set of collective attunements of a community of practice, a set that the community has harmonized around.
Those attunements can include, among other things, ways of using and reacting to words, ways of treating people, ways of looking at the world, and emotional and social attitudes. We introduced a subtype of ideologies, discriminatory ideologies, those that include attunements to in-group/outgroup distinctions, and in which members of out-groups are valued less than members of in-groups, and hence as inherently deserving of less than equal treatment or resources.
Against these background notions, here are the main points of our account of slurring: 1.
Discriminatory practice: Slurring labels the target using a negatively evaluative predicate, a slur. The label identifies the target with a disdained, despised, or hated out-group category within a discriminatory ideology, and in this sense, slurring is a discriminatory practice. 2. Presupposed ideology: Since any communicative action presupposes the ideology within which the action constitutes an exemplar of a practice, so the discriminatory ideology is presupposed by the slurring action.
3. Attention: By labeling someone using the slur category, slurring draws attention to whatever properties are stereotypically associated with that category in the discriminatory ideology. Since the discriminatory ideology coexists with an ideology that does not discriminate in the same way, performance of the discriminatory practice also draws attention to broader differences between the ideologies, and thence to the discriminatory ideology more broadly.
4. Resonances: The resonances of the slur include these attentional and emotional effects on members of the target group, a range of attunements belonging to the presupposed discriminatory ideology, a demarcation of in-group and out-group, and power differentials between these groups. 5. In-group attunement: For an in-group member, attunement to the slurring practice includes feeling such emotions as hatred, superiority, and consonance with the in-group when involved in a slurring interaction as speaker or hearer.
6. Out-group attunement: For an out-group member exposed to the slurring practice, attunement implies experience of painful loss of face, painful confrontation with the discriminatory ideology to which attention has been drawn, and further dissonance. The dissonance involves a dramatic conflict between desired private face and public face, and in many cases between a desired way of life and life as a member of an oppressed group.22 In a nutshell, to slur is to label an individual or group with a despised role within a discriminatory ideology. Within the ideology, the category is associated with distinctive stereotypical features that do not correspond to features that those outside the community of practice assigned to roughly the same set of individuals. The low valuation of these out-group members implies that to a greater or lesser extent, depending on the particular category and the particular ideology, to judge someone as belonging to the category is to judge them as despicable. In acts of slurring, slurs name a category that only exists as such within that ideology.
23 Slurring someone, then, locates them within the ideology, but it also locates the speaker.
24 Put simply, the slurring practice is performed by in-group members and names members of an out-group category, hence locating representatives of both groups in relationship to the other.
To see how the account works, we’ll first consider mild insults, and then move on to stronger slurs. Let’s start with the insult “muppet,” which is conventionalized in the United Kingdom to suggest something similar to “airhead” in the United States.
25 The resonances of “muppet” build on associations with the soft puppets that came to fame in the television shows Sesame Street and The Muppet Show, and are suggestive of someone who tends not to act or speak in an incisive, intelligent, or self-directed way. The resonances of “muppet” include not just the features that an individual sodescribed, the target, might be expected to have, but also emotions, dispositions, and cognitive attitudes.
The emotions, dispositions, and cognitive attitudes on the speaker’s side are different than those on the target’s side. Emotionally, it is clear that the speaker looks down on the intelligence of the target. The speaker is expected to feel superior, and the target inferior.
Dispositionally, the resonances include ways of using “muppet” and ways of reacting to it. A resonance of the term is that it is not simply used to insult people, although it can be, but is also used to upbraid people in a mildly affectionate manner —“You muppet!” To be attuned to the UK “muppet” practice is to use it appropriately and to react to its use appropriately, for example, not feeling cut to the bone when somebody calls you a muppet, but rather feeling a little sheepish about whatever unfortunate thing you have done without thinking properly.
The ideology associated with derogatory use of “muppet” includes various practices of using the word, but also ideas about the type of intelligent and independent behavior that can be expected of people.
Insulting uses of the term betray people’s values: the slurring use of “muppet” always belittles the target’s competence, and a harsher use, although far from being a strong slur (“Get the f- out of my way, you stupid f-ing muppet”), further betrays a strong differential in competence, and possibly power, between speaker and target.
Similar remarks apply to the Southern US use of “precious,” in phrases like “Aren’t you precious!” except that here being fully attuned to it means reacting to its use in a way that reflects the fact that you are probably not being complimented, but insulted. Southern US “precious” and other expressions that damn with praise (“Bless your heart!”) are associated with an ideology in which a distinction is made between two types of people: those who have got it together and keep things running as they ought, and a second group about whom the best we can say is that they are, for example, cute or deserving of the Good Lord’s love, but are functionally less than fully competent.
Let us consider the extent to which “muppet” and “precious” are not merely insults, but slurs.
Given our outline account of slurs, this will hold if these terms label people with disdained or despised out-group roles within discriminatory ideologies. Here we come to a point in our account that we have left open: the question of what exactly constitutes a group. One might go further than we have, and require of an out-group that it is viewed as having sufficient social cohesion that the members of the out-group form a distinctive community of practice in their own right, or that the members of the out-group jointly regard membership of the group as part of their identity. In that case, neither “muppet” nor “precious” would be slurs.
We have also suggested that, at least in paradigmatic cases of ethnic and gendered slurs, part of the attention-grabbing and revelatory power of the slur comes from contrast, a clash of ideologies that value groups differently.
It is far from clear that those using the word “muppet” are beholden to an ideology that is in sharp contrast with an alternative ideology that more fully values people whose actions betray a lack of awareness and control.
While it might be possible to define the technical term slur so as to be limited to a small set of words that everybody can agree are clear slurs, let us take a more pragmatic approach, allowing that slurs vary on a scale of strength. We suggest that a slur is stronger (a) the more reviled the target group is within the discriminatory ideology, (b) the more the group picked out by the slur itself constitutes a distinct community of practice that is central to the identity of the out-group’s members, (c) the greater the extent to which the slur use is associated with a history of oppression of the target group, (d) the greater the power of the in-group over the out-group, and (e) the stronger the contrast is between the discriminatory ideology within which the slurring practice exists and another prevalent ideology that does not devalue those in the out-group.
On this basis, the derogatory terms “muppet” and “precious” are weak slurs, perhaps so weak that some would wish to create a cutoff and say they are not slurs at all but merely members of a broader class of “pejoratives” or “insults.” The term “fat cat” is a somewhat stronger slur, because it is clearer that its use involves a clash of ideologies, although here one might say that, definitionally, out-group members are the ones with power, the oppressors rather than the oppressed. We take it that according to the ideology within which some are viewed as “fat cats,” at least some of the practices and tenets of free-market capitalism are not supported. On the other hand, we presume that some of those being labeled “fat cats” belong to a community of practice within which the practices that led to their wealth and power are acceptable. Note that in the “fat cat” case it is perfectly possible that both the ideology of the slurrer and the ideology of the slurred, the fat cats themselves, are discriminatory: the difference is that in only one of these ideologies is the target group itself reviled. We do not attempt any general taxonomy of slurs here, and leave it to the reader to consider where other slurs might fall in the spectrum. What we hope is clear is that the “nuclear bomb of racial epithets” will on this basis be a paradigmatic example of a strong slur, with (a) a high level of disdain for the target group, (b) the target group itself forming a clear community of practice that is important in shaping the identity of its members, (c) an ugly history of oppression, (d) systemic power differentials operating against the target group, and (e) a decisive clash of ideologies providing highly contrastive valuations for the target group. Such a slur is also paradigmatic in exhibiting the six properties we previously ascribed to slurs in our account: (1) use of the slur is itself a clear discriminatory practice, locating the target within a system in which they have reduced access to rights and resources, (2) the speaker presupposes a racist ideology, (3) the use of the slur draws attention to both stereotypical characteristics of the target group within the discriminatory ideology and to the discriminatory ideology itself, (4) the slur is highly resonant, emotionally, socially, and cognitively, (5) in in-group uses, speakers and hearers feel such emotions as hatred, superiority, and consonance with other racists, and (6) out-group members exposed to the slurring practice, can, if attuned to the practice, experience painful loss of face, painful confrontation with the racist ideology, and further dissonance.
Strong slurs somehow manage not merely to stick a knife into someone, but to open up old wounds, indeed an entire history of wounding behavior.
A recent paper by Elin McCready and Chris Davis offers a metaphor that is striking and helpful, even if it does, contra our view, paint a portrait of slurs as somehow magical: A sorcerer stands atop a high cliff by the sea. He raises his hands and pronounces a single word; a submerged island rises above the waves, covered with cyclopean masonry and dripping sea plants which make the precise angles of the constructions and their outlines indistinct.… The sorcerer has summoned up a city of the past from beneath the sea, where it had heretofore lain invisible. The sorcerer can do so even without knowing every detail of what lies in the city, how it is arranged, or what the consequences of calling it up will be.
Utterance of the summoning word is sufficient for the invocation. No one person observing the summoning can see all features of the conjured object. This includes the sorcerer himself.
We suggest that this case is (surprisingly) analogous to the function of slurs. The sorcerer has used a powerful word to call up a hidden, ruined city; slurs, on our view, also bring a preexisting complex of historical facts and constructed attitudes (stereotypes) about the slurred group to attention, in addition to predicating group membership in the manner of a standard nominal. However, as with the obscured nature of the summoned city, it is hard for any one person to discern exactly what those attitudes are, or what the precise historical facts being deployed are.26 How can a single word both wound and reveal an entire ideology? As Lewis Carroll’s Humpty Dumpty says (though not of slurs), “When I make a word do a lot of work like that, I always pay it extra.”27 The currency in which slurs are paid (or rather, in which they demand payment) is attention. On the view we have developed, every word is associated with a mass of resonances, essentially everything that is reliably present in contexts of use. These resonances include arbitrary features of the prevailing ideology and its constituent practices within the word’s community of practice. In general, no one community member can possibly have access to all those features, since the full set of resonances exists only in the aggregate, that is, as a collective attunement that is not identical to any individual’s attunement. For most expressions, the bulk of ideological resonances are not of great salience when the word is used. Relevant aspects of the ideological background of the interlocutors are assumed to be shared to the extent that it matters for the word to perform whatever function it is being used for.
When one uses the word “apple,” one is assuming a way of looking at the world that allows categorization according to fruit-type, but one is usually not attempting to draw attention to that background. In exchanges among adult English speakers, interlocutors using the word “apple” are part of the same relevant community of practice, with common understanding of the category. Even though this understanding is at a collective level, individuals can reasonably be expected to have similar personal attunements to the word, to apples, and to practices involving apples. So whatever differences there might be in attunements are not expected to greatly limit the ability of interlocutors to connect with each other. And if there are minor differences in attunement to apples, in most contexts this will not be emotionally or socially significant. In the terminology of chapter 6 (section 6.4), a use of the word “apple” is typically not meaningful. That is, using the word “apple” does not draw attention to the background ideology relevant to apples, because that ideology has no personal or social significance. This is where slurs differ from names for fruit. But it is precisely where slurs are similar to brand names, such as “Apple.” What makes the brand name powerful, a trademark worth protecting, is its ability to draw attention with a mere mention to a powerful array of associations— the brand identity.
What McCready and Davis illustrate is that when a slur is used, an associated ideology becomes powerfully salient, as do categorizations of group membership that are licensed by that ideology. People do not use a slur simply to help identify a particular object in the world, like identifying a piece of fruit one would like with one’s packed lunch. Rather, people use a slur to draw attention to a way of thinking about a human being. An important resonance of any slur is a high level of attention paid by interlocutors to the ideological resonances of the slur, in the sense of chapter 2 (section 2.6), and to its significance vis-à-vis group membership of the speaker and whoever is being labeled as belonging to the slur category.
For a hearer, this change in attention, like the emotional impact of the slur, is largely involuntary. Aspects of the ideology are primed by the use of the slur. In this respect, the hearer cannot help but harmonize with the speaker, in the sense that a result of the slurring act is that the interlocutors share joint attention on an ideology, even if it is an ideology one or both reject.
McCready and Davis provide an insightful way of thinking about the power of slurs. But their quote seems to suggest that slurs are remarkable, almost magical.28 To repeat, we do not think slurs are magical. Quite the opposite. Analyzing the meaning and function of slurs requires the postulation of no dedicated or extraordinary additional mechanisms within the resonance-based framework.
It might be countered that we have in fact introduced hundreds of pages of machinery. Here we can only aver that we take the machinery we have introduced to be motivated by general considerations of the nature of communication, and political language in particular. The concepts we use in accounting for the meaning of slurs—naming, presupposition, resonance, attention, and ideology—are concepts we take to be needed for the treatment of all political language, and indeed we are far from being the only ones to apply such concepts in this arena.
Slurs are the epitome of political language. They make an emotive inyour-face socially significant us-them distinction that forces people implicitly onto one side or another. In doing so, they attune people to the mores of a discriminatory ideology presupposed by the slur.
And yet we claim that slurs are grammatically unexceptional, no more complex than any descriptive name. Slurs are exceptional among descriptive names only for the way their naming depends on ideological distinctions that are themselves divisive.
The linguist Geoff Nunberg can be credited with most clearly centering the importance to understanding slurs of considering the community of practice within which they are used.29 As he memorably explains, “Here’s my thesis in a nutshell: racists don’t use slurs because they’re derogative; slurs are derogative because they’re the words that racists use.”30 Taken as a claim about sufficient and necessary conditions for slurs being derogative, Nunberg’s claim doesn’t quite hold up. Those subscribing to Nazi ideology in Hitler’s Germany can surely be described as racist, and, as Klemperer makes vivid, this community had a range of distinctive language practices.
Most of the language practices that Klemperer discussed were not slurs at all. Indeed, many of them were words or phrases that would not have stood out as racist at all, were it not for the careful way in which Klemperer characterized their relationship to broader Nazi ideology.
Nonetheless, Nunberg is right to center the derogating power of racist slurs on the properties of the people who use them, racists. At risk of watering down his pithy remark, let us just suggest a friendly addition: racial “slurs are derogative because they’re the words that racists use” to label people they take to belong to other races.31 Although some slurs are used primarily within an in-group to refer among themselves to out-group members (as in the case of slur terms for rarely seen foreign enemies), most slurs are also used for acts of derogatory name-calling, that is, slurring to someone’s face. Slurring to someone’s face differs from other acts of addressing in that its primary function is to derogate and hurt the addressee by drawing attention to characteristics listed above: a. their membership of the out-group category, and hence the location within the ideology; b. their despicability from the perspective of the in-group; c. stereotypical features of the out-group, and distinctive oppressive practices toward it.
Note that simply naming an ideology draws attention to the ideology, but is not offensive. We can refer to fascism or racism or sexism without offending anyone, unless someone thinks that by doing so we are also labeling them as having a despised ideology. Using or mentioning these terms is not (re-)enacting a practice within these ideologies, so it does not sustain the ideologies. It also does not necessarily cause anyone to lose public face, because it does not explicitly label anyone. The victim of a name-calling act of slurring loses public face, and will recognize that public loss of face to the extent that they are attuned to the utterance situation. The offensiveness and hurtfulness of a slur depends on its drawing attention to aspects of an ideology that are offensive to a person who is slurred because the ideology debases that person and normalizes oppressive practices against them.
ACCOUNTS OF SLUR MEANING Here are three views in the literature on slurs that agree with us in the claim that slurs are grammatically unexceptional.32 The first is the view that the force of a slur derives from is what is known as a conventional implicature, meaning an extra component of meaning that is conventionalized, but completely independent of the semantic contribution the word makes to truth conditions. Such a view has been defended in linguistics by Christopher Potts and in philosophy by Timothy Williamson.33 According to this view, slurs are grammatically unexceptional, since conventional implicature is an already known category of conventional meaning that does not impact assertive content. Asserting a sentence containing a slur predicates something about a group, and conventionally implicates something very negative about that group. A second view on which slurs are grammatically unexceptional is due to Christopher Hom. According to Hom, the meaning of a slur incorporates various stereotypically present characteristics, so that “fat cat” would presumably mean something like “individual who has an inappropriately large amount of money and power.” Here is his analysis of a slur word for Chinese people (which we have removed from the quote): The epithet [slur word for Chinese people] expresses a complex, socially constructed property like: ought to be subject to higher college admissions standards, and ought to be subject to exclusion from advancement to managerial positions, and …, because of being slanty-eyed, and devious, and good-at-laundering, and …, all because of being Chinese.34 In Hom’s view, the meaning of a sentence in which a slur is used to predicate a property of an individual is a gigantic conjunction of all the things that stereotypically hold of people in the slur category according to a racist ideology.
35 A third view in the literature on slurs is that their special effects result from presuppositions. Phillipe Schlenker is an example of a linguist following this line, and it is seen in work by philosophers such as Bianca Cepollaro and (in single and joint work) Manuel García-Carpintero and Teresa Marques.36 Again, presupposition is a standard grammatical notion, so this does not in itself make slurs grammatically exceptional. Let us say that a narrowly presuppositional account of slurs is one in which whatever special powers are associated with slurs or acts of slurring is claimed to derive from special presuppositions, whether presuppositions that certain things hold in the world, or, for García-Carpintero and Marques, special affective presuppositions. In that case, the account we propose is not narrowly presuppositional, because the effects of slurs in our account are not merely the results of accommodation (or rejection) of presuppositions.
We take it that many emotive and social resonances of slurs are effective resonances, that is, these resonances are conventionalized effects of slurring acts that are integral to the practices of using those slurs, and not merely aspects of the contextual background that may sometimes be accommodated by hearers. On our view, slurring may sometimes trigger accommodation of ideological background, or indeed trigger divergent accommodation against that assumed background, but at least some emotional effects are central features of the slurring practice, and not accommodated presuppositions.
There is a more fundamental respect in which not just prior presuppositional accounts but all three of the types of account distinguished above differ from ours: they assume that the meaning of a slur has a neutral core, and that there exists for each slur a neutral counterpart. Prior accounts are typically explicit about these commitments. Indeed, it might be said that all three types of analysis follow a long line of philosophical work by adopting divide-and-conquer strategies, aiming to simultaneously explain the phenomena of slurs and yet, as it were, protect the neutral truthconditional core of meaning. Thus have these prior scholars sought to conservatively maintain the important advances in truth-conditional approaches to meaning made over the last century.
Almost all prior analyses of slurs that provide detail about how the effects of slurs arise consist of a neutral truth-conditional core plus some other component that is supposed to account for how slurs are used and the emotional and social effects they have. If the extra component is not a conventional implicature or a presupposition, then it’s often some sort of pragmatic reasoning. This is the case, for example, in Hom’s analysis, in which the truth conditions of the slur are stated relative to the truth conditions of a neutral counterpart, and the emotional and social effects of the slur are explained in terms of a hearer’s ability to pragmatically infer the speaker’s negative attitude and attachment to a problematic ideology.
37 Another family of approaches not discussed above (although there are close relationships to both the conventional implicature account of Potts and to the presuppositional account of Marques and García-Carpintero) involves an expressive component of meaning; for example, the accounts of Joseph Hedger, Robin Jeshion, and Leopold Hess are of this type.38 These accounts again assume that there is a semantic component of meaning which is neutral, and that slur words share this neutral semantic core with neutral counterparts.39 While several scholars have argued against slur meanings having neutral counterparts, we know of exactly one detailed semantic analysis of the slur meaning that neither makes a commitment to there being a neutral counterpart, nor defines a slur meaning in terms of a neutral core, namely the analysis of Heather Burnett.40 It is instructive to consider Burnett’s proposal in a little more detail, since her model of meaning, drawing on work of Peter Gärdenfors and others, provides one way of cashing out what a mental space of cognitive and emotional attunements might look like, and suggests a natural path for development of our own model, although not one we seek to follow in this volume.41 Burnett focuses on the predicate “dyke” and its supposed neutral counterpart “lesbian.” Both can be used to slur, but it’s also the case that both have well-established non-slurring and even prideful uses in LGBTQ+ communities that have undoubtedly affected practices and attitudes outside those communities, although this is certainly not to claim that the result of reclamation has been the elimination of slurring usages. Burnett argues that the semantic relationship between the two terms is far more complex than one entailing the other, and furthermore that people with different ideologies may represent the semantic relationship between the two predicates differently.
In Burnett’s model, the two predicates express meanings that overlap within a complex multidimensional space, a space that includes evaluative dimensions. Since none of these dimensions are more fundamental to the meanings than any others, and it is the combination of different dimensions that jointly yields an overall affective value, there is no sense in which these meanings have a neutral core. Since neither of the two predicates are defined in terms of the other, and neither conceptual space includes the other, it cannot be said that one is the more basic counterpart of the other, even if it should happen to be the case that for some groups of speakers the words’ extensions in the world stand in a subset relation or are even identical. Furthermore, there is no assumption that either word is affectively neutral, so even if there was some sense in which one was the more basic counterpart of the other, it would not be a neutral counterpart. Thus, in this account neither term is intrinsically neutral, neither has a neutral core, and they do not stand in any simple counterpart relation, let alone a neutral counterpart relation.
There are many differences between Burnett’s account of slurs and our own, but we suppose that many of these relate to the fact that we set our account within our particular framework for the analysis of political and social aspects of communicative practice. We do not see this broader framework as inherently in tension with the main developments in Burnett’s account of slurs. A more fundamental point of contrast is found in our respective analyses of word meaning: Burnett’s model of meaning is static whereas ours is dynamic. For Burnett, the meaning of a predicate is a region of a multidimensional conceptual space, a static mental representation of a particular concept. For us, the meaning of any expression is a cluster of resonances that concern the situation prior, during, and after utterance. In this dynamic model there can be conventionalized effects of words and other expressions, for example the effect of making hearers feel good or bad. This is how we explain the exigence of slurs, their power to affect people. For Burnett any such effects must be conversational or perlocutionary effects, not built into the meaning of a slur, but generated at some point in the process of comprehending that meaning and reacting to it.
In this respect, Burnett’s model is comparable to prior accounts of slurs that either take the contribution of the slur to be purely truth-conditional, and hence non-emotive, or take there to be some expressive component (perhaps a conventional implicature), but take that component to express the speaker’s attitude. In all such analyses, unlike in our own account, the effects of slurs on hearers are taken to be indirect. It might be said that in prior accounts, including even Burnett’s, uttering a slur is like showing someone a bullet with the target’s name on. We claim that when you utter a slur you shoot the gun.
At a high level, our project in this volume of providing philosophically and linguistically satisfying analyses of political and social aspects of language led to the conclusion that the idealizations of the standard model were getting in the way, idealizations which were spelt out in the last chapter (section 9.2). It might be thought that avoiding standard idealizations would tend to make the analysis of linguistic phenomena harder. Quite to the contrary, our contention is that it is easier to analyze phenomena such as slurring, and oppressive language more generally, in a framework that, like the resonance-based framework of this volume, eschews those idealizations. For example, we think that assumptions such as those associated with the idealizations of Neutrality and Language Homogeneity and Social Homogeneity make it impossible to properly model slurring practices. Slurs cannot be analyzed in terms of neutral cores or neutral counterparts, and an analysis of slurs that assumes by default a fixed ideology among all speakers has to wrestle with nonexistent problems like the question of how it can be that the meaning of a standard socioethnic group term is both the same and different from the meaning of a slur.
(Solution: the meaning of the standard socioethnic group term in one ideology cannot possibly be identical with the meaning of a slur in another, since they are completely different practices embedded within different systems of attunement, and for similar reasons we should expect the two predicates to have different resonances even within an ideology, even if both have slurring uses.) We claim that once slurs are analyzed absent such idealizations, it becomes clear that they are grammatically unexceptional, and that slurring involves a mixture of linguistic practices that are each individually commonplace. This presents us with a challenge: how can the stark effects of slurs be explained without postulating special mechanisms? At the very least, a theory of slurs must account for three properties. Slurs are 1. ideologically revelatory, bringing a “complex of historical facts and constructed attitudes … to attention” even though “it is hard … to discern exactly what those attitudes are, or what the precise historical facts being deployed are,” as McCready and Davis put it; 2. exigent, forcibly impacting hearers’ emotional states; 3. hyperprojective, blurring the distinction between use and meaning.42 The reason slurs have become, rightly, a focus of scholarly interest is that they highlight conventional properties of language that are obscured by standard frameworks. Slurs make salient features of language that the standard model obscures, features that are also present, but less obviously so, with words that are not slurs. Like the pain we feel when kicked, what people experience as they absorb a slur has an immediacy and automaticity.
A kick has the effect it has largely because of innate properties of the body.
In contrast, attunement to the resonances of a slur is learned, rather than being a natural proclivity. In becoming attuned to a slur, people learn not only what a good dictionary might say the word means, but also ways of reacting to it. The experience of being labeled with a slur, including the emotional reaction and the dispositional reaction (e.g., a tendency to shrink into oneself, a tendency to fight back, or a tendency to attempt to educate) is something that develops over time. We will now show in more detail how the properties of slurs are to be explained within our resonance-based framework.
IDEOLOGICAL REVELATORINESS The revelatoriness of slurs consists in the fact that the use of a slur draws attention to a nexus of practices, attitudes, and emotions that distinguish between an in-group and an out-group, in short, to a discriminatory ideology. The revelatoriness of slurs is unsurprising in the resonance framework we have set out, since, in a sense, every practice is at least potentially revelatory. Every practice presupposes the ideology of the community of practice that uses it.
The revelatoriness of slurs is distinctive for two reasons. First, the role that slurs name, or use to name, does not exist outside of a very specific ideology, an ideology that typically exists in contrast to other ideologies, such as the broader, less discriminatory ideology of a larger community of practice that does not engage actively in the slurring practice. The ideologies differ precisely in the way they discriminate, that is, in the different way they separate out social categories and the different cognitive, emotional, and dispositional attunements associated with those categories.
Slurring, as we have said, draws attention to the role of an individual within a discriminatory ideology. As a consequence, slurring also draws attention to whatever features are recognized as distinguishing between ideologies.
These features are not merely resonances of the slur, but resonances that become salient when the slur is used. This is what we understand Davis and McCready to mean when they describe a slur as invoking a wealth of ideology, of bringing it into view. Even though every practice has ideological resonances, slurs are special because they bring those ideological resonances into sharp relief due to the salience of other communities of practice that do not share the ideology (presumably usually including communities that are targeted by the slur).
People vary in their individual attunements, so they are individually attuned, or second-order attuned, to various aspects of each of the ideologies to different extents. It is thus no surprise that, as Davis and McCready say, “it is hard … to discern exactly what those attitudes are, or what the precise historical facts being deployed are.” Individuals are not expected to have full access to the collective attunements of even the community of practice they are part of, since, as we argued in chapter 2 (section 2.4), collective attunement does not imply universal individual or uniform attunement. This is a point where the notion of collective attunement crucially differs from more standard notions of common ground, which typically demand homogeneity of attitude. The expectation of full individual attunement to an ideology is even lower when we are talking about people outside the community of practice that shares that ideology, that is, when we are largely talking about second-order attunements. People who are not part of the actively slurring community of practice cannot be expected to be fully second-order attuned to the ideology of the slurring community. What follows from these considerations is that an individual outside of the slurring community of practice who encounters a use of the slur can be expected to recognize that there is a mass of ways that the categorization associated with the slur differs from their own categories, and a mass of ways in which the discriminatory practices of the slurring group differ from their own practices, and yet not be able to say for certain exactly which attunements are crucial to the slurring ideology. Thus, the revelatory nature of slurring is paradoxical. The slur has revealed something dramatic, but it is not clear exactly what it consists in. The act conjures an intricate ruined city, and although it is hard to make out many of the individual taverns, steeples, and battlements, what comes suddenly into view is the central square. We see a soldier stamping on a face, a face perhaps like the addressee’s. Here lies a type of hustle. It is unclear to the addressee what the details of the oppressive structure are. And indeed, perhaps it is unclear to the speaker, who might be just a child repeating a word they heard their parents use. Slurs are transparently bad acts, but the detailed intention of the speaker, exactly what they mean by saying of someone that they belong to the slur category, is never completely transparent, and may be totally opaque. Yet here we should note that our externalist view of meaning as resonance, whereby the locus of meaning is not the individual speaker but the community of practice, means that this type of hustle is common to all communicative practices: when someone says that a particular tree is a beech rather than an elm, their intentions can only be understood relative to a broader community of practice and the perspective that the community brings on particular practices of tree-labeling. Thus, this type of hustle is not particular to slurs.
More generally, it is not peculiar to slurs that they betray the existence of an “ancient city” of ideological associations. Let us here compare McCready and Davis’s use of the ancient city metaphor to Wittgenstein’s in the Investigations: Ask yourself whether our language is complete;—whether it was so before the symbolism of chemistry and the notation of the infinitesimal calculus were incorporated in it; for these are, so to speak, suburbs of our language. (And how many houses or streets does it take before a town begins to be a town?) Our language can be seen as an ancient city: a maze of little streets and squares, of old and new houses, and of houses with additions from various periods; and this surrounded by a multitude of new boroughs with straight regular streets and uniform houses.43 The view Wittgenstein expresses is akin to the view Rahel Jaeggi puts in terms of practices existing within a nexus of other practices, as discussed in section 2.3. What is missing from this passage of Wittgenstein’s is any mention of a sorcerer. For the ideological revelatoriness of slurs is not magical but commonplace. The revelatory act does not involve conjuring up a complex nexus of practices, but drawing attention to it. Every language practice exists within and presupposes such a nexus. The practices, along with networks of supporting attunements, are there already, in the discriminatory ideology of a known community of practice. The ideological revelation offered by a slur consists merely in the fact that, due to contrast with an alternative ideology, the slur bathes the “ancient city” in the light of public attention. It turns out it was never submerged beneath the dark waters of a deep ocean to which some might have wished it banished.
Indeed, the ancient city is not so distant, and not a complete city at all. It is simply a dangerous neighborhood that many find ugly, just across the tracks.
EXIGENCE In section 2.6, we made a distinction between the instrumental power that communicative practices grant to discourse participants, and the exigent power associated with the practice itself, a power intrinsic to the practice and hard to resist, independently of the intention of the speaker. Slurs can serve as instruments of power, and are often wielded with the intention of establishing or reinforcing power, but it is the exigent power of slurs we focus on because slurs can harm even when not intended to. We have already seen at least one example of this type, involving the comedian Bill Maher. Although Maher is a famously provocative figure, and it might well be said that he knew what he was doing in choosing the slur that he used, what he overtly did was mockingly label himself using the slur, rather than directly slurring anybody else. Despite this being a self-directed use of the slur, and despite (plausibly, at least) no conscious intention to derogate Black people, the result was Ice Cube’s “knife stabbing you.” Speakers can cause pain “even if they don’t mean to.” Prior literature discusses both the immediate harms caused by slurs and other hate speech, and longer-term negative effects. At the longer-term end, Lynne Tirrell writes, “Few ask how language can make us actually sick. The key is to see how speech can generate toxic stress.”44 Mari Matsuda, though commenting on a wider range of hate speech than slurs alone, writes, The negative effects of hate messages are real and immediate for the victims.
Victims of vicious hate propaganda have experienced physiological symptoms and emotional distress ranging from fear in the gut, rapid pulse rate and difficulty in breathing, nightmares, posttraumatic stress disorder, hypertension, psychosis, and suicide.
Professor Patricia Williams has called the blow of racist messages “spirit murder” in recognition of the psychic destruction victims experience.45 In a legally oriented discussion of the need to remedy for the harms of hate speech, Charles Lawrence discusses a wide range of victim experiences, including his own as a teenager. He reports on the vivid account of a student of his facing a gay slur on a subway train, and who found “himself in a state of semi-shock, nauseous, dizzy, unable to muster the witty, sarcastic, articulate rejoinder he was accustomed to making,” to which Lawrence adds that “it is a nearly impossible burden to bear when one encounters hateful speech face-to-face.”46 Mihaela Popa-Wyatt and Jeremy Wyatt reference the same passage from Lawrence in arguing that “targets are affected as after a threat of physical assault.”47 While we are sympathetic to the main thrust of their arguments, let us note in passing that we differ from Popa-Wyatt and Wyatt as regards their analysis of the wounding effects of slurs as perlocutionary, which suggests that they are pragmatically mediated effects of slurring actions rather than being conventionalized as one of the primary functions of the action. We are not sure what sort of argument or evidence would distinguish cleanly between illocutionary and perlocutionary effects, but the distinction is anyway not significant within our resonance-based framework. If an effect is regularly associated with an action of a given type, then it is a resonance of that action. Wounding is regularly present when slurs are used; hence wounding is a resonance of slurring, a part of the practice.
Here is Jyoti Rao, an academic psychotherapist, arguing against academics even mentioning powerful slurs, and placing their use in a clinical context with further supporting citation from the psychoanalytic literature: My personal choice to avoid speaking or writing slurs in any context, including this paper, stems in part from my witnessing the intense degree of psychic harm suffered in my patients as a result of these words. The clinical literature, spanning decades, is full of examples of adult and child patients from a wide range of marginalized identities who have sustained psychic lacerations resulting from epithets directed toward them.… Slurs, like guns or whips or grenades, are designed to cause damage. Even when apparently brandished for another purpose, their original function is always nearby. 48 By focusing on the effects of slurs on hearers, and on a social community more broadly, rather than on the intention of the speaker, we invert the priorities of much of the philosophical and linguistic literature.
We do this despite building on many insights in that literature. The theoretical development of the resonance framework in this book built from the very beginning on the idea that language could be expressive. We share with the literature following Kaplan the idea that an important facet of the meaning of slurs is their expressivity.
49 Much of that literature does consider the hearer and is explicit about the offense that slurs cause by virtue of what is expressed. However, we differ from most prior literature in focusing not on the use of a slur for expression of a speaker’s feelings, but on the impression the slur makes on hearers, for this is equally a resonance of the slur.
Resonances are not the sole province of a speaker, and the exigent power of the slur itself is a power associated with the practice itself. That power is inherited from the community of practice that created the slur and imbued it with power through use. The slurring act is a manifestation of the power of a discriminatory community of practice. So, an important characteristic of the resonance of a slur is that it replicates power relationships: an act of slurring can induce a local power relationship between the name-caller and the target that, at least in the case of strong slurs, mimics a global power relationship between the oppressive in-group and the oppressed out-group.
To say that a slur has power is not yet to describe the mechanism by which that power has an effect, any more than describing a weapon as powerful would be explaining how it functions. A knife pierces because it is sharp, but how can a slur pierce? The exigent power of slurs derives from the fact that is hard to resist (i) attention-grabbing effects, (ii) awareness of what is depicted by the slur and the ideology that has been brought to attention, (iii) understanding of the speaker’s own relationship to that depiction, (iv) awareness of loss of public face, (v) negative self-image and/or tension between self-image and public face, and (vi) a concomitant feeling of dissonance. Thus, the slur resonates with the victim in the sense of section 3.9 and is meaningful to them in the sense of section 6.4, although, not, unfortunately, in a positive way.
For people who are highly attuned to the slurring practice, typically because they have been regularly exposed to it and understand its history, no reflection is needed to recognize loss of public face and to feel shame, humiliation, or other painful discomfort. They have become attuned not only to the attention-grabbing effect of the slurs, but also to the painful resonances of the practice. Being fully attuned to the practice implies feeling that pain when you are the victim. If you do not feel the pain, it is either because you are not the target of the slur, or because you are not fully, one might say not bodily, attuned to the practice. If you do not feel the pain of a slur, you may have become inured to the practice or learned to resist it emotionally, so that you no longer have the first-order attunement. The exigent power does not imply total irresistibility.
It just implies that for those fully (first-order) perceptually attuned to the practice, both attentional and emotional reactions will tend to occur nondeliberatively.
It is possible for the speaker and the victim in a name-calling act of slurring, and for relevant audience members in both the in- and out-groups, to share the same discriminatory ideology. In this case, the victim will see themselves as intrinsically less valued. Colonial power structures are designed to enforce this sense of inferiority through education and subjugation. (We speculate that in such a case the revelatory effects of slurs will be different than in cases where salient ideologies are in clear contrast.) However, a bifurcated community of practice is common. In that case, there are two relevant communities of practice, a larger community of practice (e.g., all English speakers) that contains the smaller slurring community (e.g., racist English speakers). The larger community is collectively dispositionally attuned to the slurring practice, since nonslurring community members encounter its use. Nonslurring community members may be perceptually attuned to the slurring practice, emotionally attuned to the practice, and have some mixture of first- and second-order attunements to the slurring ideology. In such bifurcated communities of practice, some degree of censure for the slurring practice will tend to arise, as those who are not attuned to the discriminatory ideology attempt to limit it.
Let us also pause to note here that while we emphasize the power of the slur itself, we also accept, with Popa-Wyatt and Wyatt, that part of the function of a slur is to establish a certain power relationship between the speaker and members of the slurred group. They go further, suggesting that the attractiveness of the power thus manifested can act as an advertisement for the discriminatory ideology: By slurring the bigot shows others the power they can acquire. On the power theory outlined here, the bigot is not talking about power, they are demonstrating power. The speaker acquires discourse power. We posit that this is emotionally appealing to audience members—who are not members of the target group—who feel less powerful than they would like. They see that they can accrue power to themselves by using a slur. Thus, a perlocutionary effect is to make others desire the power the bigot has grabbed. Increasing desire is different to increasing acceptability. Both are required for audience members to join the side of the bigot.50 HYPERPROJECTIVITY We discussed the notions of projectivity and hyperprojectivity in section 5.4. We take the hyperprojectivity of slurs to consist in the fact that mere mentions of some slurs, including quotative uses, can cause offense. Here the N-word is the clearest case.
Here is Jyoti Rao, again, discussing situations in which the N-word is mentioned as an object of study, or quotatively, at academic conferences: What happens after the slur is uttered has been equally consistent, and equally notable. Some small part of the group, typically comprising the few people of color and others present who come from marginalized backgrounds, attempt to bring attention to what they have just experienced. Often, they express palpable pain, clearly expressing the destabilizing effect of hearing a word strongly associated with white supremacy spoken at a conference. Aside from occupying a marginalized social identity, those speaking are frequently earlier in their career; hold positions as graduate students, junior faculty, or analytic candidates; are less financially secure; and are speaking to people with greater institutional, organizational, and other forms of power. In response to hearing from these participants, the user of the word, and several other white people in the group, seem mobilized to counter what has been reported about the consequences of the epithet’s use. They begin to explain that speaking the n-word is benign, even salutary, and advocate for why the word should be used freely by white people and psychoanalysts.
In all the cases I have seen, the person using the word, as well as the people roused in support, have appeared unmoved, unreceptive, and unapologetic, even when it is repeatedly pointed out to them that their speech has caused harm.51 Mere mentions of a slur have the potential to do harm. Depending on the slur, mere mentions can help perpetuate marginalization of oppressed groups and solidify oppressive practices. As Judith Butler puts it after describing slurs as “badges of degradation,” mentions of slurs “unwittingly recirculate that degradation.”52 Quotation marks play a complex role in discourse and metadiscourse, but whatever quotation marks do, they do not “neutralize.” We are not even sure what that could mean. The idea that quoted mentions of slurs are neutral is pernicious.
Note here that the philosophy and linguistics literature is peppered with slurs, though practices of mentioning slurs have evolved in recent decades, as have broader societal norms. Those quoting or mentioning slurs in this literature are not being intentionally vindictive, and many are explicit about their policies on slur mention and the reasons for it.
But we think that there is still a great deal of unnecessary explicit mentioning of slurs, and that this might be in part the result of an error centering on more-or-less blind acceptance of a certain interpretation of the use-mention distinction, an idealization at the heart of contemporary analytic study of meaning.
According to this idealization, merely mentioned or quoted language is in some sense inert, contributing only its form to the content of what is said, and not directly contributing its meaning. It is as if quotes or italics were a lead-lined box, so the scholar who uses these devices cannot possibly be risking harm with the radioactive material inside. But theoretical prejudice should not interfere with reality. So let us just say this. How can the scholar know that their belief that a quoted slur will cause no harm is correct? If some evidence should arise suggesting that someone was harmed by exposure to a mere mention, would that not suggest that maybe it would be better to treat at least certain slurs a little more carefully, perhaps being very sure that there is no reasonable alternative to mentioning them before doing so?
Let us be clear here. Our analysis of hyperprojectivity does not imply that no slur should ever be mentioned in scholarly work, or indeed that there is any slur at all that should never be mentioned in scholarly work. We recognize that in general there could be many reasons to mention particular slurs, for example legal, lexicographic, or historical, or within a process of reclamation, and we are in no position to legislate the appropriacy of such mentions. It is also important to see that our account does not imply that quoted mentions of slurs will have exactly the same power or effects as other utterances containing slurs, such as simple non-quotational acts of name-calling. The hyperprojectivity of slurs consists in some of the resonances of clear acts of slurring being present to some extent as resonances of communicative acts that involve mentions of slurs, not in the resonances of the two types of speech act being identical. So, while there is danger in the mere mention of a slur, that does not mean that mentioning is the same as slurring.
Let us suppose that X is a strong slur which is in current usage. If someone uses X in a name-calling act, for example saying to someone “You are a <intensifier> X!” then they are canonically instantiating a slurring practice. Likewise, an in-group use like “I blame those <intensifier> Xs!” would be a canonical instantiation of a slurring practice. In such paradigmatic cases, the resonances of the slurring act descend in a clear way from the extension of the slurring practice, that is, the history of prior usages, and hearers who are attuned to the practice will tend to be affected by the utterance in accord with those resonances.
Now let’s go to the opposite extreme: an entry in a comprehensive general dictionary for a language in which a slur is simply listed alongside a definition. This is paradigmatic as a lexicographic practice, but not a paradigmatic act of slurring. The lexicographic entry can still succeed in drawing attention to both the despised role described by the slur, and to its ideology, and indeed it would be a poor dictionary if it did not do that. Further, it is possible that some people will be hurt by seeing the entry in the dictionary, since their attention will have been drawn to something painful, and that still others will be offended by the dictionary entry.
Certainly, someone could use the dictionary entry in order to perform an act that while not canonical could be readily recognized as a slurring act, for example by mailing the relevant dictionary page to someone who people holding to the slur ideology would take to belong to the target group, perhaps with a ring drawn around the slur, thus making it clear that they took the slur to label the recipient. But that merely shows that a dictionary entry can be abused, can be exploited in an act of slurring, and does not make the production of the dictionary entry itself into a slurring act. To the extent that a lexicographic entry is sufficiently distinct from the extension of the practice of slurring that it is not taken to instantiate that practice, it will not affect people attuned to the practice in the same way as a paradigmatic act of slurring would. The slur will have at most the power to draw attention to something painful, and thereby potentially cause some pain or discomfort, and not the exigent power of a slurring act.
Many cases of slur quotation in the philosophical and linguistic literature on slurs fall in between these two extremes. In such academic contexts, it is clear that the author would deny that they are performing a slurring act, and we will not take issue with this. What we would say is this: first, while one occurrence of a slur (like in a dictionary) will draw a certain amount of attention to the slur category and its attendant ideology, and potentially cause pain, repeated uses will draw a lot of attention to them, and potentially cause significantly more pain. This pain may be magnified if a reader takes some of the repetitions to be gratuitous, which might suggest to them that the writer is repeating them with disregard for the feelings of readers, and hence is also performing something just a little closer to a canonical act of slurring. Second, when scholars report on real or hypothetical contextualized acts of slurring, they may sometimes be demanding of the reader an act of imagination, the imagining of an act of slurring. Suppose that some reader is attuned to the slur, and feels they could plausibly be a target of such a slur. No scholar outside the target group can be relied on to say whether in such a case we should expect the reader to successfully compartmentalize, merely imagining what the pain of encountering the example would be, or whether in such a case a reader might actually feel pain, feel targeted, and perhaps be affected in other ways, like being silenced or having a decreased sense of wanting to be part of that academic milieu.
That slurs are hyperprojective remains a controversial position in philosophy and linguistics.
Recently, a special journal issue appeared with the title “The Challenge from Non-Derogatory Uses of Slurs”; the editors write that although “what is peculiar to these expressions is their so-called hyper-projectivity,” the collection “focuses on how slurs can be used in non-derogatory ways.” They go on to say that “it is disputed whether slurs are derogatory when they occur in reported speech” and that “most scholars agree that quotation marks can seal the derogatory force of slurs.”53 If they are right, then on this issue we find ourselves in sharp disagreement with most scholars. Quoted slurs can not only derogate and give offense, but can cause harm.54 (We will turn to the scholars the editors identify as not agreeing that that quotation marks can seal the derogatory force of slurs, namely Luvell Anderson and Ernie Lepore, below. Unfortunately, we are not in agreement with their account either, though for quite different reasons.) Theories of slurs set within variants of the standard model generally do not do well at explaining hyperprojectivity of the sort discussed in this section. If the effects of slurs are to be explained as conversational implicatures, then there is a general problem that conversational implicatures typically don’t project. If the effects of slurs are to be analyzed by analogy with expressives like “oops” and “ouch,” as discussed in chapter 2, then we arrive at the problem that it is completely unclear what is expressed when an expressive is quoted. Presuppositional accounts and conventional implicature accounts, like those discussed earlier in this chapter, make some headway, since there are analyses of projection phenomena for these types of meaning.
Presuppositions, as standardly analyzed, are projective but cancellable (as in the well-worn “The King of France is not bald: there is no King of France,” in which the presupposed existence of a French King is canceled). Yet it is hard for extant presuppositional accounts to explain why some presuppositions should be more projective than others. The effects of slurs are hyperprojective and not easily cancelable. This requires a kind of special pleading for most presuppositional theories.55 Even if, like Bianca Cepollaro, one has a theory that treats slurs as presuppositions that are not capable of being contextually cancelled, it remains mysterious why even quoting slurs should be objectionable, or how it could possibly be a source of harm.56 While it is well established that conventional implicatures tend to be more projective than most presuppositions, neither conventional implicature accounts nor prior presuppositional accounts can directly explain why mere mentions should lead to any projection at all.57 Again, special pleading would be needed to explain this property, for example a theory of quotation that treated different conventional implicatures differently.
While the details of exactly what effects project depend on the strength of the slur and associated considerations of ideology, the simple fact that some effects of slurs are hyperprojective can be explained straightforwardly in terms of attention. Drawing attention to something does not require use, at least not in the sense of the standard use-mention distinction in philosophy of language. The mere mention of a word suffices to draw attention to its meaning, and often to its reference.
The primary point of names is to draw attention to individuals.
Descriptive names like vocative uses of the appellations “Professor” or “Private” simultaneously draw attention both to a person and to a role. The attention that names draw to individuals, one could say, is hyperprojective.
Even if one uses a name within quotation marks, the audience’s attention is still drawn to its referent. If in a report on an important news event, say concerning revelations of spying, one paragraph includes a quote from an anonymous official that mentions you by name, you are sure to soon find reporters at your door. By merely mentioning you, the report has drawn attention to you. If someone mentions the combination of a safe in a quote, for example, “I heard Marie muttering the number 1673,” they have drawn attention to the named number, which may or may not have security implications.
Perhaps this notion of drawing attention seems mysterious, worthy of special philosophical consideration before being centered in a theory. We don’t deny that attention is worth special philosophical attention, but the notion of attention we require is no more mysterious than that evoked, sometimes using the term “salience,” in much theorizing about meaning, especially in the context of work on anaphora. Consider this example: 1. My friend said “I met Trump in New Hampshire!” but I don’t believe she met him there.
We take it that this constructed example is unobjectionably acceptable English. But note that two of the pronouns in the last clause, “him” and “there,” refer anaphorically to entities introduced in the quotation. Entities can only be referred to using pronouns if they are salient. Therefore, something must have drawn attention to them. The names clearly did that. This attention-drawing effect is no more cut off by surrounding quotation marks than light is cut off by clear glass.
Slurs are often said to be epithets, the term “racial epithet” being used almost synonymously with “racial slur,” and acts of slurring are often described as name-calling, which is also invariably negative. If a teacher merely refers to Johnny Smith as “Johnny Smith,” Johnny can’t legitimately say the teacher was “calling him names.” It’s clear that slurs are commonly thought of as name-like. However, our arguments regarding attention do not depend on slurs being names per se. It is a completely general fact that a use of a predicate inside a quotation draws sufficient attention to the kind or property it denotes that this kind or property has increased availability for later anaphoric reference. Thus in (2) the quoted mention of a pet dodo makes dodos salient, and the later pronouns “they” and “one” refer to dodos at a kind level. Neither quotation marks nor the act of quotation veils the attention drawn to kinds by the quoted material.
2. Johnny is so cute. This morning he said: “Guess what pet I have? A dodo!” I couldn’t tell if he doesn’t know they’re supposed to be extinct, but let’s not tell him we know he doesn’t have one.
Slurs, then, are both like names and like any other predicate, in drawing attention to whatever individuals, properties, or kinds they make explicit.
Making things explicit implies drawing attention to them. The way in which attentional processes are modulated by knowledge of language is undoubtedly complex, but we take it as uncontroversial that once something has become salient, its salience will tend to decay only slowly, presumably as a result of basic psychological processes that are common to humanity, if not to all higher animals. The key to understanding the hyperprojectivity of slurs is recognizing that hyperprojectivity is not in and of itself in the least unusual. What most distinguishes the hyperprojectivity of slurs from the hyperprojectivity associated with most other classes of predicates is simply that people care about it.
Grammatically, slurs are like other predicates, though they are predicates that, like “professor” and “private,” have a tendency to be used as epithets, and indeed as vocatives.
A practice of slurring, like any other, occurs within a community of practice. Like all predicates that are associated with negative evaluations within an ideology, slurs are used insultingly, to draw attention to despised characteristics of an individual.
The crucial way in which slurs differ from other predicates is that the kind they draw attention to is a construct within a distinctive, discriminatory ideology, an ideology that makes in-group/out-group distinctions such that “members of out-groups are valued less than members of in-groups, and hence as inherently deserving of less than equal treatment or resources” (cf.
section 2.6).
The use-mention distinction has been central to the development of logic and analytic philosophy over the last century, but assuming the distinction to be sharp is an unwarranted idealization, and unhelpful in many cases. Let’s consider a few. First, a letter that says “I hereby apologize for saying ‘It’s the one-year anniversary of my getting the position that you so badly wanted’: I shouldn’t have reminded you of it.” is not a convincing apology. Second, the use-mention distinction is unhelpful for distinguishing between whether an improvisational jazz musician has merely quoted a few bars of a classic or has actually played it. Third, and more pertinently, the use-mention distinction is unhelpful for distinguishing between whether someone reproducing an image of a threatening racist display hanging from a tree in a back yard has merely exhibited someone else’s display or has themselves performed a racist act. Finally, we submit that the use-mention distinction is at best a blunt instrument for distinguishing between different occurrences of slurs. That is because both using a predicate and mentioning a predicate can draw attention to whatever kind or property the predicate denotes, and drawing attention to something is one of the main functions of slurs. What makes the hyperprojectivity of slurs significant is that part of the function of a slur is to do just what many quoted or otherwise embedded slurs do, namely draw attention to the disdained or reviled role of a group within a distinctive discriminatory ideology. It follows that a quoted slur can function somewhat like a non-quoted slur.
Let us briefly head off a potential objection to our suggestion that attention is primary. Some might point out that when something is mentioned, the Gricean maxim of relevance can then explain the communicative significance of the mention. On this view, when a hearer encounters a mention of a term, they must then reconstruct the intention of the speaker in mentioning whatever it is. The reason why a friend keeps mentioning figure-skating news, the hearer might reason, is because he wants people to think about his recent figure-skating success. We completely agree that such reasoning processes occur, and that they can explain aspects of the communicative significance of mentions, including mentions of slurs.
Joe Biden once caricatured former New York City Mayor Rudolph Guiliani by saying, “There’s only three things he mentions in a sentence—a noun, a verb, and 9/11.”58 What he was referencing was Giuliani’s habit of drawing attention to the terrorist destruction of the World Trade Center, and hence his role in the city’s recovery. Although it might be argued that Giuliani was simply obsessed by the event, it seems more plausible that, as Biden was implicitly suggesting, Giuliani was strategically mentioning 9/11. Donald Trump has a habit of beginning sentences with “People say … ” and related locutions. The discursive logic of this strategy is multifaceted, but at least part of it is that it allows Trump to draw attention to ideas without taking responsibility for them. An authoritarian leader (and perhaps any politician) needs to be a master of attention, and often the strategic point of their utterances is to draw attention to something.
However, as an explanation of the direct emotive and attentional effects of slurs, such a strategic explanation would be backward. Relevance does not explain attentional effects. It’s rather the case that a premise of relevance-based argumentation is that the speaker has drawn attention to something. The question the Gricean theorist is then asking is, Why has the speaker drawn attention to it? The question we are asking is, Given that the speaker has drawn attention to something, what effects will that have independently of what the speaker’s intention was? We do not dispute that there are effects that should be explained in terms of the speaker’s intention, or indeed that such reasoning is important in considering what form of counterspeech is appropriate (e.g., blame and censure vs. education). What we dispute is that considerations of the speaker’s motivations are needed to explain the exigent power of the slur, its hyperprojectivity, or its ideological revelatoriness. And our explanation for this hyperprojectivity centers on a claim that we take to be somewhat self-evident: the attention-drawing power of a construction is not plugged by quotation, or, for that matter, by other metalinguistic embeddings.
THE ORDINARINESS OF SLURS If language is treated primarily as a vehicle for conveying information and information is supposed to consist in neutral and objective facts about the world, then the tools and resources one constructs will be focused in a way that makes expressive and exigent language mysterious. As a result, slurs will be mysterious, as their central function is expressive and exigent, not descriptive. Slurs appear as a puzzle because of the assumptions embedded into the ideal-language model (the predicate calculus, for example, lacks slurs).
We have urged a reorientation of the evidence base of the theory of meaning toward political speech. In so doing, expressive and exigent properties of language use emerge as equally central to descriptive properties. We use language to describe, but also to insult, bond, and emote.
In the vocabulary of our picture, words resonate with things, with social identity, with emotions, with practices, with values, and with much else. We do not think of one of these resonances as “primary” and the others as derivative.
From our perspective, slurs are not mysterious, but rather a kind of paradigm case. If there is such a thing as ordinary language, then slurs are ordinary. Words belong to speech practices, in most cases, multiple speech practices. But a slur is a characteristic expression of a certain kind of speech practice, one that is part of an ideology that negatively stereotypes the targeted group. That does not make slurs special from the point of view of the linguist. It just makes them the right sort of thing to be used in practices of labeling and insulting, as well as further practices of commanding, prohibiting, punishing, or blaming. All of these further practices have linguistic dimensions, but their existence does not need to be explained by a special theory of the linguistic properties of slurs. Slurs are just about as ordinary as language gets. But then again, in line with the discussion of section 10.1, to claim of anything that it is ordinary is perhaps to say as much about your own ideology as about the thing you so describe.
In their introduction to a seminal set of essays on critical race theory in 1993, the editors (Mari Matsuda, Charles Lawrence, Richard Delgado, and Kimberlé Williams Crenshaw) discuss the defacement of a poster of Beethoven in Ujamaa, Stanford’s Black-themed house, which two White freshmen had represented as a caricature of a Black man, and upon which they had scrawled the N-word. The editors write, The power of the poster’s message was derived from its historical and cultural context, from the background of minstrel shows, of racist theories about brain size and gene pools and biblical ancestors that has shaped our conscious and unconscious beliefs about the intellectual capacity of Blacks.59 The defaced poster harmed because it was an endorsement of an ideology of White superiority and Black subordination. It is the connection of the Nword to this ideology, the connection of the caricatured image of a Black man to a past history of minstrel shows, that gave the poster its dangerous power.
In a similar vein, Judith Butler writes, Clearly, injurious names have a history, one that is invoked and reconsolidated at the moment of utterance, but not explicitly told.
This is not simply a history of how they have been used, in what contexts, and for what purposes; it is the way such histories are installed and arrested in and by the name. The name has, thus, a historicity, what might be understood as the history which has become internal to a name, has come to constitute the contemporary meaning of a name: the sedimentation of its usages as they have become part of the very name, a sedimentation, a repetition that congeals, that gives the name its force.… If we understand the force of the name to be an effect of its historicity, then that force is not the mere causal effect of an inflicted blow, but works in part through an encoded memory or a trauma, one that lives in language and is carried by language.60 We agree with Butler’s description of both the conventional and the perlocutionary facts involved with the use of slur terms, and our apparatus, as it were, cashes her idea out in our own currency. The historicity of a slur is its embeddedness in a speech practice, which in turn is part of an ideology. When one performs an act of name-calling using a slur, one is making salient an ideology that subordinates them, calling it to their explicit attention, and applying it to them. The reason such name-calling is injurious is because it is the characteristic expression of an ideology that places that person in a subordinate role by virtue of their membership in a targeted group.
Crucially, the ideology associated with a slur term is conventionally connected to it: it has (as Butler puts it, above) “come to constitute the contemporary meaning of a name.” In our terms, a slur is a characteristic expression of such an ideology. Using the slur functions to manifest that ideology. In this sense, the ideology and the slur are interwoven—the slur and its resonances are quite literally part of that ideology.
But it is not just slurs that are connected with ideologies. It is a moral of our book that words generally are connected with ideologies. The word “wife” is connected with an ideology, an ideology about gender. Eric Swanson is right to argue that “slurs cue ideologies”; but it is also true that “wife,” “mother,” “secretary,” and “boss” cue ideologies, albeit that we sometimes miss the cue.61 Philosophers recognize that there is a kind of overlap in what one wants to say about slurs, and what one wants to say about words that are not slurs.
Here is Robin Jeshion, clearly recognizing that the normal word “janitor” also regularly carries with it an ideology, even a subordinating one: It is useful to contrast the identifying component of slurs with uses of language that suggest or signal lower status in-a-role. For example, remarks like “He is a janitor,” designating someone’s occupation may be used as a put-down by signaling lower occupational status by virtue of being a janitor. But such signaling is highly contextual, no part of the semantics of “janitor.” This is markedly different from slurs, for which the identifying component is encoded in every context of literal use.
Moreover, even in contexts in which “janitor” is used to signal lower status in a role, they fail to negatively evaluate the targets vis-à-vis their humanity, construed along a moral dimension—qua person. “He’s a janitor, and the finest person I know” is perfectly acceptable; substitute a slur for “janitor,” and it reads as highly problematic.62 Here, Jeshion distinguishes between a slur, which signals lower status, according to her, via its semantics, and a word like “janitor,” which, by her lights, can be used to signal lower status, though this is not, according to her, part of its semantics. It’s useful to lay out our analysis of the similarities and differences between a slur and a word like “janitor,” to compare and contrast them with Jeshion’s.
We do not accept Jeshion’s claim that the felicity of “He’s a janitor, and the finest person I know” removes the imputation of an ideology that ranks janitors as lower status than other professions, any more than “She’s a secretary, and the finest person I know” is free of such imputations (it isn’t).
Low social status is consistent with being a “fine person”—as Kate Manne has emphasized in her work on gender, accepting one’s lower social status is in fact often how one comes to be regarded as a fine person (in her analysis, a fine woman).63 It’s probably difficult to rid any use of “janitor” (or “secretary”) of its association with an ideology that accords it a lower social status than other professions.
Even if “janitor” only sometimes brought with it an imputation of lower social status to those who occupy that role, it would not, in our framework, follow that this imputation was not conventional. Each use of a term manifests a speech practice, and simultaneously presupposes that very speech practice. If there are contexts in which “janitor” does not signal lower status in a role, that could be both because resonances are inherently probabilistic, allowing for a mixture of uses in different types of context, and because “janitor” belongs to several distinct speech practices, only some of which involve ideologies that rank janitors as having a lower status than other professions. When “janitor” is used in a context in which it is part of an ideology that ranks janitors as having a low social status because of their role, the ideology functions in communication exactly as the ideology associated with a slur functions when the slur is used. In both cases, the ideologies are presupposed. Even if “janitor” is not always used to manifest an ideology that ranks janitors as having a lower social status, there is, on our view, no sharp difference in linguistic kind between “janitor” and slurs, beyond the greater tendency for slurs to be used vocatively, and the fact that “janitor” is unlikely to be reclaimed.
In Jeshion’s work, we find a rich account of the perlocutionary effects of slurring terms, effects that, for her, are “exclusively pragmatic,” and do not obtain “by virtue of conventional linguistic properties of slurring terms, i.e., meanings or conventional rules of use.”64 We find much to agree with in Jeshion’s explanation of the perlocutionary effects of slurring uses. For example, Jeshion attributes the difference in force and harmfulness between a use of the N-word and a use of “honky,” a slur for White people, to the histories behind the ideologies associated with these words; as she writes, “The former occurs against the background of current widespread racism, history of slavery, and historical civil rights struggles for AfricanAmericans, and nothing comparable for Caucasians.”65 We agree with Jeshion’s analysis, which builds on the crucial insights of critical race theory with which we began this section. But, with Judith Butler, we insist that there is a conventional tie between standard uses of these words and these different histories.66 Jeshion’s view about slur terms, which we take to be windows into the conventions associated with them, also explains the fallacy of conflating the distinct notions of taboo and slur. In German, words associated with Nazi ideology, such as Lebensraum, Führer, and Rasse, are taboo. They are taboo because of their connections to horrific historical practices. But these words are not slurs. Conversely, there are slurs, like “fat cat,” which are not taboo, presumably in part because they are not connected to a history of horrific treatment (in this case, of rich Americans). Some slurs are taboo and some taboo terms are slurs, but the topic of what is a slur and the topic of what is a taboo are entirely distinct subject matters. Words that are connected to particularly horrific histories and practices typically can neither be used nor mentioned without raising those histories and practice to salience.
In a series of papers, Luvell Anderson, Ernie Lepore, and Matthew Stone develop a “prohibitionist” account of slurs, explaining the various properties slurs have, including hyperprojectivity, as resulting from the existence of taboo.67 Leaving aside for the moment the question of whether this counts as an explanation (one might worry that the “explanation” is that it is offensive to mention slurs because they are unmentionable without causing offense), or whether it can possibly explain which specific resonances project and which don’t, there is something right about it. The fact that something is taboo means that mentioning it will immediately catch listeners’ attention, and indeed lead to the suspicion that the purpose of mentioning it was to call it to the listeners’ attention. Perversely, the social development of taboos around slurs strengthens their effectiveness, by increasing their attention-drawing power. However, the development of these taboos does not so much explain anything about slurs, as add to the complexity of generalizing. For a taboo is itself an idiosyncratic complex of practices. These practices include counterspeech, censorship, and education, all of which are applied in different ways for different slurs, and for which their application varies over time. What have been deemed acceptable mentions of various racial or gender-related slurs has evolved continuously during our own lifetimes. The non-taboo nature of relatively weak slurs like “fat cat” and pejoratives like “muppet” is illustrative of a problem with using taboo as the basis of an explanation, rather than something that itself needs to be theorized, because the taboos relevant to even the paradigmatic cases of racial slurs are complex. It would be at the very least a gross oversimplification to maintain that in the reconstruction-era South the Nword was taboo, when an officer of the law could openly use it as a term of address. It was a word in general use in a powerful community of practice.
And it was a powerful word. How can the power of a word, its exigence as we have termed it, be explained by the illicitness of using it, if those with power feel its use is licit? Similarly, the revelatoriness of slurs, their ability to bring to the fore an ugly discriminatory ideology, is not explained by their taboo nature, but is rather largely responsible for the imposition of societal taboos. Last, the hyperprojectivity of slurs is not explained by their taboo nature.
Rather, hyperprojectivity is part of what makes the taboos against using some slurs so sweeping.
Imagine someone explaining why you shouldn’t murder by saying “Well, it’s illegal.” Murdering and many other violent acts are illegal, but that is not satisfying as a general explanation for why you shouldn’t perform them, for it merely pushes us back to the question of why there are such prohibitions. Further, in the case of physically violent crimes, nobody would claim that the harmful effects of the acts result from their prohibition. So it is with hurtful methods of social stigmatization more generally: they are hurtful independently of any prohibition against such stigmatization. In particular, slurs are not injurious because they are prohibited; they are prohibited because they are injurious.
Some theories of slurs, like Jeshion’s, add an expressive component— for Jeshion, as with other expressivist theories, it is part of the semantics of slurs that they communicate emotion, in this case contempt, for their targets. We agree. And yet, on our way of looking at things, the fact that uses of slurs attune their hearers to the speaker’s emotional attitude toward their targets does not make them in any sense distinctive. Affective resonance is a general conventional feature of words, not unique to slurs.
Calling someone “a mom” can carry warm resonances of family values, and simultaneously a negative evaluation of their status in industrial society.
Talking about something growing in the garden as a “weed” or a “plant” carries emotional resonances that connect with its perceived value.
The supposedly distinctive linguistic behavior of slurs is, for the most part, a straightforward consequence of the fact that they are part of language, embedded in histories, practices, and ideologies, whose nature explains how their capacities to harm have become conventionalized.
Because words, generally, carry such histories, we must take care not to treat slurs as special or distinctive in this regard. So doing perpetuates a myth—that “ordinary language” is not linked to ideologies in just the same way. By focusing on slurs, we are led away from thinking critically about the ideological weight of perfectly ordinary terms. Slurs are not distinctive in being the vehicle of ideology. Language generally is a vehicle for ideology.
There are nevertheless important and interesting differences between slurs and other terms.
Using words like “wife,” “mother,” “boss,” and “janitor” presupposes ideologies that provide people with social roles. But these ideologies are not normally highlighted by the use of these words.
When we describe someone with their use, we presuppose these ideologies. When someone says “Mom!” we do not normally expect any relevant clash of attunements between interlocutors as to the role of being a “mom,” and while attention is drawn to the role of the addressee, we assume that what the utterance usually does most strongly is draw attention to the person rather than her role. A male politician who calls someone “my wife and the mother of my children” is presupposing a gender ideology. But he is not usually doing so as the explicit and manifest point of his utterance. Using language like this can be a kind of hustle. If a politician intentionally uses this phrase to signal his allegiance to his ideology, but wants his audience to think it was incidental, that is hustle.
In contrast, slurs do not hustle in this way, at least in the mouths of competent language users. Slurs overtly draw attention to a role that is a point of contention; those who do not subscribe to the slur’s associated ideology might deny that the role exists at all. Ideologies are constituted by practices, including speech practices. Slurs are characteristic expressions of certain ideologies. Having such an ideology about a group, one can manifest the ideology with that word. Slurs thus present their ideologies (although not the details of those ideologies) explicitly. The main point of a use of a slur is to give expression to the ideology for which it is a characteristic expression, and to do so openly. Slurs evoke ideologies by naming despised roles; part of their point is to draw attention to those ideologies and to draw attention to the role those labeled by the slur play within that ideology. Because slurs name ideologies, they bring the erasure of ideologies by ideal-theoretic models especially saliently to attention.
According to Jessie Munton, prejudice involves undue attention to certain properties of a group, and it hinders us from gaining important knowledge.68 The idea that what matters in prejudice is not, or at least not only, what claims are made, but what is drawn attention to is an important one. Although she does not discuss slurs explicitly (at least in recent work with which we are familiar), it’s clear that slurring is an expression of prejudice, and thus that her account should apply to slurs. According to Munton, then, a slur could potentially be a vehicle for prejudice not because it falsely describes a group, but because it unduly draws attention primarily to negatively evaluated properties of the group.
While we agree with Munton that attention is highly relevant to the way people conceptualize the groups around them, let us note a tension arising from the fact that Munton focuses on undue attention. It might be argued that some terms characteristically express ideologies about an identity group, ideologies that highlight negative properties of that group, but nevertheless may help us think better about social reality. For example, the term “Karen” is used to express something negative about White women.
But, arguably, “Karen” may help orient people to features of social reality, and thus may be part of an ideology that better reveals the world, rather than masking it. Is “Karen” still a slur?
The word “prejudice” is negative, implying distorted evaluation or unfair behavior. One might choose to characterize slurs as words that are characteristic expressions of flawed ideologies, or prejudices in Munton’s sense, or one might define slurs in such a way that at least some of them focus appropriately on negative characteristics of a group, say because those negative characteristics are societally problematic and deserve special attention.
Whether one defines slurs so that they always imply prejudice in Munton’s sense, or allows that slurs can convey a justifiable bias against a group is a largely terminological question, the answer to which we do not need to legislate here.69 So-called reclaimed uses of slurs are ones in which members of the targeted group “reclaim” the slur, to perform a different speech act than slurring. As Luvell Anderson discusses, Black Americans typically use the N-word to address one another, rather than to slur or insult one another. 70 The reclaimed use of slurs is a mystery for many theories, especially those that make slurs into a special or distinctive category. But from our perspective, there is nothing surprising about slur reclamation. Reclaimed slurs are used as a bonding mechanism between those who are traditionally targeted by the slur. To slur is to label someone with an out-group role, so one thing that is clear is that when someone who appears to be part of that group uses the slur word, they are either denying their membership of the slurred group, or else are not performing an act of slurring. Reclamation concerns a case where someone does not deny their membership of the group, but rather denies the power of others over the nature of the category.
Only if we are close can we talk freely, so the more potentially painful the ideology a term invokes is, the better its invocation is as a means of testing and intensifying our closeness.
Reclaiming slurs can also be a way of challenging the very ideologies that the slur has been part of. And here again, there is no contradiction, but rather a confirmation of the slur’s power. There are few better ways of manifesting power and stubborn opposition than to hold in your hand your enemy’s weapon. The more powerful the weapon, the better: why wave your enemy’s toothpicks in front of them, if instead you can fire off an artillery round?
We can focus on “reclaimed uses of slurs” as if these are some baffling linguistic phenomena, but that is a sign of a misunderstanding of the phenomena. Reclaimed slurs are one manifestation of a general tendency, one that we also see when Jewish people make Jewish jokes with one another, or even jokes about the Nazi Holocaust (jokes that do not typically even involve reclaimed slurs). Bonding, group strength, and pride are often gained through subversion—waving in the face of a history of oppression by others weapons they fashioned but no longer own. One characteristic reclamatory use of slurs involves embedding the slurs into a new speech practice, a practice of bonding, which is a practice one can only engage in as a group member.
71 Robin Dembroff has argued that the category gender-queer is what they call a critical gender kind.
72 On their analysis, self-describing as “gender-queer” is a way of undermining the ideology of the Western gender binary.
Slur reclamation shares this destabilizing function—it is a method of undermining the problematic ideology behind the slur. Slur reclamation belongs in a category of methods used to attack and undermine ideologies historically associated with terms. Bonding with a slur directed against one’s group has the effect of undermining the negative ideology associated with the slur. Not all reclaimed slur uses involve bonding—as Luvell Anderson (p.c.) has pointed out to us, Cornel West’s use of “thugs and gangsters” to describe Western Imperialism is not a bonding use of “thugs and gangsters.” But it does embody the function of Dembroff’s critical gender kinds—it undermines an ideology.
It is not accidental that slurs have attracted so much attention in philosophy and linguistics in recent years. Focusing on slurs brings out the connection between language and ideology, a connection less salient with so-called “ordinary” words. Slurs highlight this connection.
When slur terms are used to slur a target directly, in acts of namecalling, they are attempts to fit the audience into an ideology that subordinates them. More generally, slurring, especially openly, and with terms with direct associations with violent ideologies, has characteristic effects on the social context that reinforce hierarchies of value and worth.
In July 2020, US Congresswoman Alexandria Ocasio-Cortez was accosted by her colleague Representative Ted Yoho, who labeled her using, among other things, a highly charged negative gendered slur. Representative Ocasio-Cortez, in a speech on July 23, 2020, addressed the effects of such language directly on the House floor: Now what I am here to say is that this harm that Mr.
Yoho levied, … tried to levy against me, was not just an incident directed at me, but when you do that to any woman, what Mr. Yoho did was give permission to other men to do that to his daughters. In using that language in front of the press, he gave permission to use that language against his wife, his daughters, women in his community, and I am here to stand up to say that is not acceptable. I do not care what your views are. It does not matter how much I disagree or how much it incenses me or how much I feel that people are dehumanizing others. I will not do that myself. I will not allow people to change and create hatred in our hearts.73 As Ocasio-Cortez here points out, slurring a woman as she was slurred “gives permission” to a misogynist ideology. Since slurs are the characteristic expression of various toxic ideologies—in this case, a misogynist one—using a slur openly makes it permissible to voice that ideology. It normalizes that ideology.
Ocasio-Cortez’s speech closely echoes the analysis given by Mary Kate McGowan of oppressive speech. Focusing on a very similar misogynist discourse involving the same slur for women, McGowan writes that such overt misogyny changes the norms in the conversational context, making it subsequently “conversationally permissible to degrade women.”74 Slur words are a particularly powerful way of normalizing ideologies.
Using a slur in a speech act of slurring is an overt manifestation of the negative ideology associated with a slur; slurring the target group is a part of a speech practice constitutive of that ideology. And it is the very overtness of a slur in an act of slurring that gives it its ideological power.
Eric Swanson has been particularly clear about the attentional function of slurs. As Swanson writes, Uses of slurs strengthen ideologies because the use of a slur makes manifest the speaker’s consent to and endorsement of an ideology, encouraging the speaker and others to feel that their own consent to and endorsement of that ideology would not be out of place.75 Slurring using a word that is unambiguously a slur does not covertly smuggle in adherence to a discriminatory ideology, but manifests that adherence. Therefore, performances of such acts by competent members of a community of practice may come close to an ideal of straight talk, the intended effects of the act being transparent to hearers.
By making manifest what others would prefer remained hidden, slurs give permission to openly endorse the ideology they presuppose. Clearly, much straight talk is far from ideal.
10.3. Genocidal Speech Genocidal speech is speech that targets a social group and provides justification for its genocide. Genocidal speech is connected to the most extreme formation of an antagonistic ideological social group (section 6.6), a community of practice whose discriminatory ideology is structured around a strongly negative collective emotional, dispositional, and attitudinal attunement to another group. Genocidal speech builds the strongly negative collective emotional attunement by defining the group as being those whose existence is most directly and existentially threatened by the supposed enemy. A genocidally antagonistic ideological social group is a community of practice whose identity is based on being existentially imperiled by the existence of another group.
Genocidal identities are formed by representing the target group as an existential threat.
One way the existential threat is posed is as what Susan Benesch and her collaborators in the Dangerous Speech Project call “accusation in a mirror,” “attributing to one’s enemies the very acts of violence the speaker hopes to commit against them.”76 We began this book with Cleon’s speech in the Mytilenean debate, the classical example of demagoguery in Western history. Cleon argues for slaughtering all the Mytilenean citizens, on the grounds that they would slaughter all the Athenians if the tables were turned. Cleon’s speech is an exercise in genocidal speech. Its principal narrative structure is accusation in the mirror.
Antiquity’s paradigm of demagoguery was genocidal speech. It is a central case.
Genocidal speech inherits its power from lengthy histories of conflict, oppression, and revenge, and this is why the presence of such a history is one of the five parameters in our definition of the strength of a slur, in the last section. (It’s condition c.) In “Genocidal Language Games,” Lynne Tirrell describes the presence of such a history as “social embeddedness”: Derogatory terms are most effective when they are connected to networks of oppression and discrimination, with the weight of history and social censure behind it. This is what most clearly marks deeply derogatory terms from other sorts of slurs. Let’s call this the social embeddedness condition. Social context, with embedded practices and conventions, is the major source of the power of derogatory terms that are used to dominate, demean, or dehumanize people.77 If, out of the blue, a politician in the United States starts describing AngloSaxon men as “vermin,” it will not have much effect. Similarly, if, out of the blue, someone on social media describes a New York Times columnist as a “bedbug,” this too should be expected to have little effect, although predicting the short-term effects of any insult is impossible.78 To resonate meaningfully, the description of the target must take place in an environment in which there is a deep history of conflict with the targeted group, and the term fits in the right way into the context of a historical narrative. The derogatory speech must connect the targeted group in the right way with this richer history of conflict. Here we recall both Deva Woodly’s discussion of the importance of history and social context to resonance, discussed in chapter 2, and Rahel Jaeggi’s broader analysis of practices as located within a nexus; we might say that the derogatory practice is supported and gains its own strength by virtue of being one thread within a fabric, but the fabric is itself no more than the interweaving formed by the threading of individual practices. Social embeddedness consists in the presence of such a fabric, but the practices in question are oppressive, and the fabric they form is a discriminatory ideology. Given the presence of this ideology and its attendant discriminatory practices and attitudes, the derogatory term produces consonance for oppressors and dissonance for the target group, that is, it resonates positively for the oppressors and negatively for the target, becoming meaningful for both, and developing its own exigent power through the context into which it is woven.
What we observed of slurs in the last section is true of derogatory language more generally, and is true of the speech that attends genocides.
Derogatory language is not particularly extraordinary in respect of presupposing ideology, of depending on a network of other practices and attitudes, but neither is genocidal language that is not directly derogatory.
The highly charged phrase “Heil Hitler” is an honorific expression, and derogates indirectly by manifesting the power of a social group with a highly genocidal discriminatory ideology. It is an obvious example of a practice that was associated with a genocide, but which, despite its oppressive power, is not what we would usually think of as a slur. But again, the mere fact that it is socially embedded, meaningful in the context of a particular ideology, is not in itself extraordinary. All linguistic practices are like this. The ability of oppressive language to resonate positively or negatively depends on the fact that the term draws attention to particular roles within a discriminatory ideology, and this attentional change is crucial to the development of extreme consonance or dissonance. In Rwanda, the postcolonial situation created a division between Hutu and Tutsi, riven with jealousy, anger, and competition over favored status with the colonial occupiers. In Rwanda, poisonous snakes are a muchdespised threat, and killing them with machetes an honor.
Hutus were told that Tutsis were enemies as snakes were enemies. This fit into a history of conflict between Hutu and Tutsi and made possible communicating to Hutus that being Hutu meant, in part, to target Tutsis for elimination via a violent practice typically directed toward snakes.
The power of the “super-predator” campaign in the United States in the 1990s, directed against young Black men, derives similarly from social embeddedness. In the United States, there is a long history of demonizing Black men by connecting them to horrendous violent crimes, a process that Khalil Gibran Muhammad has called “writing crime into race.”79 Regularly, US elections take place in a backdrop of panic about “Black crime.” Many Americans operate in this social context and use words that are part of speech practices that legitimize the racist ideology that Black American men tend to be criminal by nature. The introduction of a word like “superpredator” must be understood in this long historical context.
It was a novel addition to anti-Black racial ideology, and it was effective not primarily because of its isolated shock-value, but because of what it added to that ideology. A racist ideology is one that includes attunements to racist practices, attitudes, and affects, one that has a downstream effect on attention to behaviors. The term “super-predator” added a grenade to an already existing racist ideology, which acted like its grenade launcher.
Let’s illustrate the central concepts of the study of genocidal speech with a contemporary example of genocidal speech, employed to justify Russia’s 2022 unprovoked invasion of Ukraine. On April 3, 2022, the Russian official press agency, RIA Novosti, published an article titled “What Should Russia Do with Ukraine?”80 The historian Timothy Snyder has aptly described the article as “Russia’s genocide handbook,” noting “The Russian handbook is one of the most openly genocidal documents I have ever seen.”81 Snyder is a preeminent historian of mass killing. Snyder’s assessment means that this document is one of the most openly explicit examples of genocidal speech that has ever been written. From the outset of Russia’s invasion of Ukraine, Putin gave as a justification the “denazification” of Ukraine.
The document fleshes out this justification.
It begins by describing Ukraine as “the enemy of Russia and a tool of the West used to destroy Russia.” It begins, therefore, with an accusation in the mirror—the claim that Ukraine will do to Russia what Russia intends to do, via its invasion, to Ukraine. It proceeds to develop the logic behind the accusation. The West has supposedly abandoned its traditional European values in favor of an ideology described as “Western totalitarianism, the imposed programs of civilizational degradation and disintegration, the mechanisms of subjugation under the superpower of the West and the United States.” Russia is “the last authority in protecting and preserving those values of historical Europe (the Old World) that deserve to preserve and that the West ultimately abandoned.” The supposed Western destruction of Russia by its tool, Ukraine, is an existential threat that requires nothing short of a genocidal response.
In his 1935 speech “Communism with the Mask Off,” Nazi propaganda minister Joseph Goebbels describes the threat of Bolshevism in similar terms, though more explicitly antsemitic: “In its final consequences it signifies the destruction of all the commercial, social, political, and cultural achievements of Western Europe, in favour of a deracinated and nomadic international cabal which has found its representation in Judaism.”82 Just as the RIA Novosti article represents Russia as the protector of the West’s traditional values in the face of the “Western totalitarianism” that seeks to annihilate them, Goebbels represents Nazism as the protector of the West’s traditional values against the existential threat to civilization posed by Judaism.
The “genocide handbook” document outlines a history of grave historical wrongs Russia has supposedly suffered at the hands of the West.
“Russia did everything possible to save the West,” it proclaims, yet “the West decided to take revenge on Russia for the help that it had selflessly provided.” Ukraine is represented as the primary tool of the West’s treachery toward Russia. It is a call to do to Ukraine what Ukraine supposedly is doing to Russia and traditional values on behalf of the West: destroy it.
According to the document, “Ukronazism” is the ideology that defines Ukraine as an independent nation. It is a version of Nazism, but far worse: “Ukronazism poses a much bigger threat to the world and Russia than the Hitler version of German Nazism.” It defines Ukrainian identity as an “antiRussian construct that has no civilizational substance of its own”—that is, the central feature of Ukrainian identity is its antagonism to Russia, and it has no other nature. This means that “unlike, for example, Georgia or the Baltic States, history has proved it impossible for Ukraine to exist as a nation-state, and any attempts to ‘build’ such a nation-state naturally lead to Nazism.” The document describes at length the practices that constitute “denazification” of Ukraine. They include “mass investigations” to uncover personal responsibility for “the spread of Nazi ideology” (Ukrainian democracy) and “support for the Nazi regime” (an independent Ukraine).
The punishments are described as forced labor, the death penalty, and imprisonment. The practice of denazification involves “the seizure of educational materials and the prohibition of educational programs at all levels that contain Nazi ideological guidelines” (i.e., the prohibition of anything mentioning Ukrainian identity).
The document focuses on the historical role of Russia in fighting the West. It is an attempt to provide a new definition of Russian identity, to transform the meaning of being Russian. It defines Russians as a genocidally antagonistic ideological social group. To be Russian is to accept the logic of genocide, the accusation that Russia is existentially threatened by Ukraine, the West’s tool of war against Russia. This provides a narrative frame within which pro-Putin nationalist Russians can develop collective harmony. Within this frame, to be Russian is to be motivated by the logic of genocide to engage in the total annihilation of Ukraine. The document represents the practice of “denazification” of Ukraine as the purest exhibition of Russian identity. According to its logic, Russian identity is most perfectly exemplified by brutal and violent revenge against the treacherous people who willingly allowed themselves to be the West’s tool in its mission to destroy Russia.
The propaganda used to motivate Russia’s 2014 and 2022 invasions of Ukraine was not created out of whole cloth. It was socially embedded in Russia’s long history of justifying colonial brutality in Ukraine.83 It includes a history of representing Ukrainian identity as a fake anti-Russian construct.
It is a campaign that Russian President Vladimir Putin dramatically accelerated in the years before the 2022 invasion, starting well before he ordered annexation of Crimea in 2014. It is only with this background that Putin’s propaganda of “denazification” of Ukraine had purchase and power to motivate genocide.
Ideologies are collective attunements to, among other things, practices of behavior, and this includes verbal behavior. Introducing a word to describe a group without a richer background ideology about that group will have little effect, as a one-word description of a group is hardly rich enough to serve as an ideology, much less a racist one. A discriminatory ideology consists of a web of attunements to practices, attitudes, and affects toward members of the targeted group. To take an earlier example, simply calling members of a group “vermin” out of the blue does not create such an ideology. It is an idle naming practice. In contrast, adding to a preexisting racist ideology the practice of calling members of that group “vermin” or “super-predators” will be a mobilizing factor in violent action against them.
Similarly, “denazification,” far from being created out of whole cloth, is woven tightly into an existing fabric, a nexus of practices.
In the enterprise of genocide, words like “vermin,” “snakes,” and “traitors” have a certain function—they mark those labeled for violent action, often mass death. In order for these words to fulfill this function, they must be connected to violent practices via an ideology. For example, in Rwanda, snakes were associated with the practice of being killed by machetes, including in ceremonial circumstances that granted manhood to boys. Introducing these new labels only will have force if there is an already existing ideology into which they harmonize.
Building an ideology that justifies violence against a group takes time—a mere label such as “vermin,” without such a background, will be powerless; a one-word speech practice will not harmonize in isolation and is likely simply to be shrugged off by the audience that is meant to absorb it.
One of the key concepts in the study of genocidal speech is that of existential threat. The power of genocidal speech to mobilize depends on local histories of conflict and hatred, which undergird the power of accusations of existential threat. Addressing the legacies of the histories of conflict and hatred that give genocidal speech its power is part of diminishing its effectiveness. Similarly, to undermine the effectiveness of the vocabulary of “super-predators,” one must weaken or eliminate the antiBlack racist ideology in the context of which “super-predators” gained the power to eliminate empathy toward Black juveniles. To undermine the genocidal ideology motivating Russia’s 2022 invasion of Ukraine, one must make the character of modern Ukraine vivid, showing that it is not an “antiRussian construct,” but an ordinary nation-state in the positive sense, which developed in a way reminiscent of other European nations, with its own language and traditions, not just as a construct defined in political and cultural opposition to Russia. This kind of work, uprooting ideologies that enable genocidal speech to mobilize toward violent goals, is core to critical theory’s antifascist practice.
Accusation in a mirror is a way to present a group as an existential threat. Great Replacement Theory (GRT) is another way to cast a group as an existential threat. GRT presupposes an ideology of the nation, according to which it is historically ethnically and religiously homogeneous and pure, with a set of fixed traditions, typically including patriarchal ones. In GRT, this ideology, including its traditions and their practitioners, is supposedly existentially imperiled by a target, usually an influx of foreign races, ethnicities, or religions. The foreign elements are sometimes described in the narrative of GRT as vermin or diseases. This influx is presented as an existential threat to the nation, its purity, its traditions, and its values. GRT is employed as a justification for mass violence against these foreign elements, as well as their internal agents.
GRT was central to the official Nazi motivation for the genocide of the Jews of Europe.84 In Mein Kampf, Adolf Hitler discusses the version of GRT that underlies National Socialism, which has as its center Jews, who engineer laws to allow mass non-White immigration. Hitler writes, “The infection of the blood, which hundreds of thousands of our people undertook as though blind, is, … promoted by the Jew today. Systematically these black parasites of the nations ravish our innocent young blond-haired girls and thus destroy something that can no longer be replaced in this world.”85 Hitler here suggests that Jews are organizing a mass replacement of the Aryan population of Germany by non-Whites.
Mussolini’s rhetoric in the run up to Italy’s colonial war against Ethiopia in 1935 harped on racial paranoias about the decline and replacement of the White race. In 1934, Mussolini published a front-page newspaper article “The Death of the White Race?” and in the final paragraph posed as a key political issue: “It is a question of knowing whether in the face of the progress in number and expansion of the yellow and black races, the civilization of the white man is destined to perish.”86 This text laid the ground for the racism and segregation imposed by Italians during the war against Ethiopia in 1935 and later the racist and antisemitic laws of 1938.
In the United States, the fantasy of racial replacement goes back centuries. In 1892, Ida B.
Wells, in “Southern Horrors,” traces the justification of the racial terror of lynching to the White male horror at the prospect of White women having children as a result of consensual relationships with Black men.
GRT has been central to the mass violence of fascist regimes, such as the National Socialists and the Italian fascists. It has also been central in the United States to racial fascism. GRT is the ideology of the Ku Klux Klan.
Federico Finchelstein and Jason Stanley have suggested that the connection between GRT and fascism, both European and American, is not accidental.
GRT is central to fascist ideology.
87 This raises the question of whether the ideology of GRT, like fascism, is inconsistent with democracy.
Madison Grant’s highly influential 1916 book, The Passing of the Great Race, focused on the replacement of Whites in America by intermingling with Black people, as well as with immigrants, such as “Polish Jews.”88 All of these groups were considered by Grant to be existential threats to Nordic Americans, the most important of America’s “native class.” (Grant, it should be noted, was fine with the presence of Black people in America, as long as they played a subordinate role.) Grant’s book was an exercise in scientific racism, arguing that “Nordic whites” are superior intellectually, culturally, and morally.
Democracy is a system based around two values: freedom and equality.
Fascists use GRT to argue that these democratic ideals are existential threats to the nation.89 The first democratic ideal, equality, brings with it a demand for racial equality, which violates the racial hierarchies presupposed by GRT. The second democratic ideal, freedom, is a threat to fascist purity and tradition. For example, LGBTQ freedom allows violations of traditional patriarchal norms. Freedom threatens the ideology of a pure, patriarchal nation that is presupposed by GRT. The very first chapter of Grant’s book is “Race and Democracy,” in which he contends that democracy is a threat to Nordic supremacy, because democracy leads inevitably to greater immigration and equality between races.90 There is a close connection between fascism and genocide. The connection is not merely historical, that is, due to the prominence of fascist regimes that have committed genocide (Nazi Germany) or racial terror (the Ku Klux Klan). The connection is conceptual. Great replacement theory is both a kind of genocidal speech and a kind of fascist speech. GRT can be used to justify physical genocide, and it can be used as an argument for the necessity of a fascist regime. GRT also figures in cultural genocide. The ideology presupposed by GRT involves a fixed set of traditions, ones that could be threatened, say, by foreign religions. GRT has been central to the arguments of European far right-wing parties against non-White and non-Christian immigration. It is present in the United States in Christian Nationalist movements, who defend Christian nativism against the threat of non-Christian immigration.
In GRT, equality is taken to be a tool central to the replacement of the dominant ethnic group. In GRT, freedom is regarded as a tool used to destabilize tradition. It is thus of note that GRT is behind two propaganda campaigns sweeping the politics of multiple democratic nations during the writing of this book: the campaign against Critical Race Theory (CRT) and the campaign against gender ideology.
CRT, pioneered by academics such as Kimberlé Crenshaw and Derrick Bell, is an approach to understanding persisting gaps in the United States between Black and White Americans, for example, in wealth, housing, incarceration, and education. CRT rests on the fact that America’s institutions were designed initially by those who sought to preserve their own status and power, including by maintaining racial hierarchies privileging White Americans.
CRT holds that while attitudes may have changed, the practices and structures they left behind persist (for example, in residential housing segregation). The political will to implement the massive structural change needed to overcome structural inequality of this sort—in education, law, finance, etc.—has not yet emerged. First, as conservative commentator David French has noted, “Time and again, there are non-racist reasons for wanting to maintain the structures racists created.”91 Secondly, the forces that zealously protect these structures as a way of preserving wealth and power remain powerful in American life.
CRT shows that, with a background of stark racial inequality, practices that are on their face neutral or meritocratic can function instead to reinforce disparities. For example, in America, public schools are to a large degree funded by local taxes. Because of racial segregation, many Black Americans are caught in underresourced schools, which leave them at a temporary educational disadvantage, harming them in meritocratic “raceblind” academic competitions.
So much for CRT. What is the campaign against CRT, and upon what is it based? The goal of the campaign is allegedly to ban the teaching of CRT in schools. But if one bans explaining to students that race-blind principles have been behind racist agendas throughout US history, one is in fact banning US history. That is the real goal of this campaign.
The effectiveness of the anti-CRT campaign rests on the discovery that the expression “critical race theory” can be used, like “welfare,” as a political weapon. The words “critical race theory” tap into a long-standing racist narrative, the version of great replacement theory known as White replacement theory. According to White replacement theory, the struggle for Black equality is really an attempt, masterminded by Marxists (historically identified as Jewish) to grab power by replacing the culture and political power of White Americans. This narrative is embedded into US history. As a result, it has a certain intuitive familiarity that lends itself to collective harmony within a large segment of the US population.
In other words, the basis of the anti-CRT campaign is a version of great replacement theory.
Its power to mobilize is inherited from the embedded narrative power of White replacement theory. The campaign against CRT consists of fomenting panic about the political and cultural power of White Americans being replaced by Black Americans. Just as Russian propaganda represents Ukraine as the vehicle for the West’s plan to destroy Russia’s traditions and identity, CRT is presented as the Marxists’ tool to destroy White traditions and identity. Its effectiveness is aided by the fact that White replacement theory narratives have a deeply ingrained history in the United States.
Great Replacement thinking is also behind another propaganda campaign central to the far right, not just in America but worldwide. This is the propaganda campaign against “gender ideology.” The campaign against gender ideology, as Elizabeth Corredor has shown, arose first in the 1990s as a countermovement to the success of the feminist and LGBTQ movements.92 The feminist and LGBTQ movements had challenged dominant biological understandings of gender, instead understanding gender as a cultural construction. The countermovement, which first started in the Catholic Church but has now spread worldwide, has taken the form of representing these social constructionist views of gender as an existential threat to traditions, particularly patriarchal traditions. In short, the campaign treats feminist and LGBTQ social movements as existential threats to traditional patriarchal values, which they would replace. The campaign against gender ideology is also based on great replacement theory.
Great replacement theory is, in its essence, antidemocratic. The campaigns against CRT and gender ideology inherit their antidemocratic character from their basis in great replacement theory. The basis of the antiCRT campaign is the desire to preserve White cultural and political dominance. The anti-gender ideology campaign opposes feminist ideals, like female equality, since they threaten the inegalitarian ideology of patriarchy.
Russia’s war in Ukraine is, as we have seen in detail, motivated by genocidal speech. In the campaign, all of the elements we have discussed in this section come together. Russian propaganda in support of the invasion represents Ukrainian identity as an existential threat to Russian identity, the Ukrainian language and traditions as threats to replace the Russian language and traditions, and, finally, democracy as an existential threat to Russian nationalism.
10.4. Bureaucratic Speech Oppressive language does more than represent violence; it is violence; does more than represent the limits of knowledge; it limits knowledge. Whether it is obscuring state language or the fauxlanguage of mindless media; whether it is the proud but calcified language of the academy or the commodity driven language of science; whether it is the malign language of law-without-ethics, or language designed for the estrangement of minorities, hiding its racist plunder in its literary cheek—it must be rejected, altered and exposed. It is the language that drinks blood, laps vulnerabilities, tucks its fascist boots under crinolines of respectability and patriotism as it moves relentlessly toward the bottom line and the bottomed-out mind. Sexist language, racist language, theistic language—all are typical of the policing languages of mastery, and cannot, do not permit new knowledge or encourage the mutual exchange of ideas.
—TONI MORRISON93 In her 1993 Nobel Prize address, Toni Morrison calls attention to the oppressive nature of “obscuring state language” and “the commodity driven language of science.” Morrison sees in the language of efficiency an ideology that enables mass violence and dehumanization.
We are going to call the kinds of speech that are the target of Morrison’s address bureaucratic speech. According to Morrison, bureaucratic language enables violence by masking it. Specifically, it masks the effect of policies on humanity by speaking of social reality entirely in terms of technocratic ideals. We do not have any kind of detailed theory to offer. Instead, by focusing on some examples in which bureaucratic speech does seem to be causally implicated in violence and harm to communities, we use her insights to unpack it.94 Bureaucratic speech certainly does not always harm. Much bureaucratic speech is benign, or at the very least not implicated in enabling mass violence. Morrison is calling our attention to the situations in which bureaucratic speech enables mass violence, and in distinctive ways, in virtue of its status as bureaucratic. The general challenge bureaucratic speech poses is to explain why speech that is not on the surface harmful is nevertheless almost invariably present as part of the surrounding justification of mass atrocities.
In Golfo Alexopoulos’s 2017 book about Stalin’s Gulag, she documents how “the Gulag leadership masked the violence of physical exploitation” by the use of bureaucratic terminology.
95 The central concept of Gulag bureaucracy was “labor utilization,” which was used as a yardstick of camp function, and was designed to set the Gulag prisoners in a positive light; to do this, four categories were used, which Alexopoulos lists as (i) “labor force, working in industry,” (ii) “occupied in the service and maintenance of camps and colonies,” (iii) “Sick,” and (iv) “Not working for various reasons.”96 Given what we know about the extraordinarily high rates of malnutrition and associated ills in the Gulag, a suspiciously high percentage of prisoners in the Gulag were classified in “labor force, working in industry.”97 The reason here is that in the Gulag bureaucracy, many activities counted as “industry,” and even gravely malnourished prisoners too weak to climb out of their beds, could count as participating in some industry, for example knitting. The capacity to participate in some industry or other precluded classification as sick, so prisoners in the Gulag could find themselves denied medical care until they were already well on their way to death from malnourishment and other forms of imposed weaknesses.
The word “industry,” from the official bureaucratic vocabulary, was defined in a way that masked mass starvation in the Gulag camps. Other official state vocabulary was completely explicit in its purpose, of masking harsh reality. Forced labor for sick patients was called “labor therapy.”98 These specific Gulag examples are from the language of Gulag bureaucracy, and are also paradigm examples of speech that harms by masking mass violence. In each case, the harm stems from a devious use in the bureaucratic language of some ordinary term (“industry” in one case, and “therapy” in the other). These examples suggest that bureaucracies mask mass violence by the strategic redefinition of ordinary terms, and the strategic aim of the redefinitions is to mask the violence of the state. We should expect that in states whose policies and institutions are implicated in mass violence, some of the language of its bureaucracy was strategically designed to mask it. That is what bureaucracies in such states regularly do.
In May 2022, an investigation by journalist Dean Kirby found that many thousands of Ukrainians from territories occupied by Russia during its invasion had been “sent to remote camps up to 5,500 miles from their homes.”99 Analyzing local news reports, Kirby identified up to 66 camps. In language clearly evocative of the Gulag, one activist told Kirby that the “state treats them as a labor force, as objects, moving them around without taking care of what they need.” The deep social embeddedness of the ideology of the Gulag forms the background of the Russian genocide in Ukraine today. Classifying prisoners as “willful work refusers” who, because of starvation, are too weak for work was a very deliberate attempt by the Soviet regime to mask the Gulag’s crimes. Alexopoulos’s view is that the Gulag labor utilization classification was intentionally designed to mask these crimes. In this case, it was the state bureaucracy that was the source of the intentional masking of mass violence. The reason that state bureaucracies are such effective sources of such masking is, as Toni Morrison argued, because bureaucracies are supposed to be guided by technocratic ideals, which in turn are supposed to be neutral. Institutions that are governed by supposedly neutral ideals, such as bureaucracies, are particularly effective production sites of perniciously ideological state propaganda, because they are supposedly governed by neutral technocratic ideals, which are supposed to be by their nature incompatible with the practice of using vocabulary in the service of strategic masking of injustice.
As we saw in chapter 7, claims of neutrality regularly play just this strategic role of masking the presence of problematic ideologies.
Turning to the United States, a Michigan law allowed for “emergency managers” to replace elected officials in cases of “financial emergency.” Emergency managers were supposedly experts in “financial efficiency.”100 Mayors, city councils, and school districts supposedly in “financial emergency” across Michigan received emergency managers, including the city of Flint. Under the guise of financial responsibility, the emergency manager regime delivered hazardous waste to Flint’s homes, which six thousand or so children bathed in and drank for almost eighteen months.
According to Shawna J. Lee and her University of Michigan School of Social Work based colleagues, African-Americans comprised 14 percent of Michigan’s population. Despite their much lower total representation in the Michigan population, 51 percent of African-Americans in Michigan were under an emergency manager at some point from 2008 to 2013. The Hispanic or Latino population comprised just 4.4 percent of the Michigan population, yet it too was also overrepresented, with 16.6 percent of all Michigan Hispanic or Latinos under an emergency manager at some point during 2008–13. In contrast, during the same time period, although nonHispanic Whites comprised 76.6 percent of Michigan residents, only 2.4 percent of Whites in the state were under an emergency manager. Michigan’s Emergency Manager Law is supposedly guided by neutral metrics of “financial efficiency.” Yet in practice, it found 97.6 percent of Michigan White citizens to be capable of democratic self-governance, in contrast to the 51 percent of Black citizens, who were deemed not worthy of elected representation, due to “financial emergency.” The language of financial efficiency served to mask anti-Black racism (Morrison’s “racist plunder”). The decisions made by the appointed emergency managers, most obviously in Flint, but also in the Detroit public schools, resulted in terrible harm to their populations, largely by favoring attempts to earn or save money over any other goal. The Detroit Public School System was moved into a new statewide system called the Educational Achievement Authority (EAA). In September 2014, investigative journalist Curt Guyette wrote an article titled “The EAA Exposed: An Investigative Report.”101 The article showed that the EAA seemed to function more like a product-testing laboratory for a software company, using its ten thousand children to test its product, “Buzz,” and the taxpayer money of their parents to support the company.
In Michigan, the technocratic ideal of financial efficiency masked racism and harmed the populations it was supposed to help. In most cases, the policies passed by appeal to this ideal were not even financially efficient —for example, refusing to allow municipalities under emergency managers to sue banks for fraudulent loans is clearly something done in the interest of banks. Nevertheless, what happened in Michigan under the guise of financial efficiency was not as deliberately bad faith as the Soviet classification of labor utilization.
Still, even if it had resulted in financially efficient outcomes, the consequences would be unacceptable in a democracy. And yet that very fact is masked by the language.
Bureaucratic speech attunes its audience to a limited set of allegedly neutral metrics. Such metrics by their nature are insensitive to a host of harms to human dignity and human flourishing. The function of the metrics is to direct people’s attention away from these harms.
Bureaucratic speech pushes attention away from the issues that affect them, and not toward any meaningful a way to think about those issues. As Randal Marlin says (in a discussion of what Jacques Ellul termed “rational propaganda”), “Citations of facts and figures leave the impression of great rationality, but the hearer is unable or unwilling to analyze the figures and is persuaded by the appearance of rationality rather than by coming to grips with genuine reality.”102 If one designs a school system solely around financial efficiency, one will be led to use the school system as a method to earn money by, for example, exploiting students. Exploiting students to earn money may be financially efficient, and it may lead to an expanded endowment for the school district. But it is an unacceptable outcome in a liberal democracy, as these policies have consequences that diminish human flourishing.
Bureaucratic speech is woven into the fabric of an ideology that privileges financial exigency above human rights and equality, and privileges the voices of a technocratic elite in-group who is supposedly in a privileged epistemic position above the majority. An ideology that revolves around practices in which the voices of an in-group have greater weight than those of the out-group is, according to our definition, discriminatory, even if the values expressed by practitioners are superficially democratic.
In May 2022, a US Supreme Court decision revoking Roe v. Wade, authored by Supreme Court Justice Samuel Alito, was leaked. One of the arguments Justice Alito gave for his decision involved adoption. In a footnote, Justice Alito cited the Center for Disease Control, commenting that “domestic supply of infants relinquished at birth or within the first month of life and available to be adopted had become virtually nonexistent.” Justice Alito appealed to this bureaucratic conceptualization of women as “domestic suppliers” of babies to reverse women’s access to abortion—if women’s function is as “domestic suppliers of babies,” they are hardly agents with rights.103 Almost definitionally, bureaucratic speech represents itself as neutral and nonideological. This is often given as a selling point. As we saw in chapter 7, this selling point is fictional. Claims to neutrality are invariably ideological. The supposed neutrality that bureaucratic speech typically represents itself as embodying is incoherent.
The claim to neutrality is central to the usefulness of bureaucratic speech in masking atrocities of various sorts. Calling attention to such atrocities often requires explicit moral language. As Golfo Alexopoulos notes, “Violent human exploitation constituted the essential purpose of Stalin’s Gulag.”104 The purpose was not labor utilization.
Patricia Hill Collins has drawn attention to another venue in which bureaucratic speech can be oppressive, in the academy. In describing the ideology of positivism, Collins writes, Positivist approaches aim to create scientific descriptions of reality by producing objective generalizations. Because researchers have widely differing values, experiences, and emotions, genuine science is thought to be unattainable unless all human characteristics except rationality are eliminated from the research process. By following strict methodological rules, scientists aim to distance themselves from the values, vested interests, and emotions generated by their class, race, sex, or unique situation. By decontextualizing themselves, they allegedly become detached observers and manipulators of nature.105 As Collins points out, when positivist approaches are the dominant knowledge-validation method, they rule out other sources of knowledge, such as narrative testimony. Positivist ideology masks reality when it blocks legitimate sources of knowledge about it.
Bureaucratic language is openly antidemocratic. The practices of using bureaucratic language signal membership of an exclusive community of practice, which cannot be freely joined because of systemic societal issues and active gatekeeping. The practices have resonances that include the bureaucratic ideology and the practices of wielding power within it. The use of bureaucratic language establishes collective harmony among the controllers, which alleviates individual responsibility, a point familiar from Hannah Arendt.106 The bureaucrats live in their own echo chamber, like a cult, but a cult with power. A natural analogue is leadership by religious zealots, even if the state religion happens to be technocratic economics.
Their ideology is presented as objective and inevitable, but it is in fact discriminatory and arbitrary. The maintenance of bureaucratic practices in the face of a community whose humanity has not been fully recognized involves active resistance to convergent accommodation, which would open the echo chamber to other ways of thinking. Such accommodation would signal common humanity, acceptance of the will of the people as they express that will, and membership of a joint community of practice with common goals.
Bureaucratic practices block such accommodation, leaving control in the hands of its exclusive community of practice. 1. Morrison, “The Nobel Lecture in Literature,” 200.
2. Arendt, Eichmann in Jerusalem: A Report on the Banality of Evil, 48.
3. We do not here give an analysis of oppression, but for one with which we are in sympathy, see Haslanger’s Resisting Reality: Social Construction and Social Critique (327), where she defines what it is for an individual x to be oppressed *as an F* by an institution I. Here, we can think of oppressive language itself, or certain ways of talking, as an institution that oppresses individuals for their identities.
4. Arendt, Eichmann in Jerusalem: A Report on the Banality of Evil.
5. Arendt, Eichmann in Jerusalem: A Report on the Banality of Evil, 108–9.
6. Browning, Ordinary Men: Reserve Police Battalion 101 and the Final Solution in Poland.
7. Klemperer, The Language of the Third Reich, 41–47.
8. The innocent ease with which a convinced young Nazi could use the term Strafexpedition is seen in Klemperer’s anecdote about a conversation with someone who had been a friend: “How are things with you at work?” I asked. “Very good!” he answered. “Yesterday we had a great day. There were a few shameless communists in Okrilla, so we organized a punitive expedition.”—“What did you do?”—“You know, we made them run the gauntlet of rubber truncheons, a mild dose of castor-oil, no bloodshed but very effective all the same, a proper punitive expedition in fact.” (Klemperer, The Language of the Third Reich, 43) 9. Klemperer, The Language of the Third Reich, 43.
10. Klemperer, The Language of the Third Reich, 45.
11. Klemperer comments in a nuanced way on the importation of mechanistic language, as seen in this passage: One of the foremost tensions within the LTI: whilst stressing the organic and natural growth it is at the same time swamped by mechanistic expressions and insensitive to the stylistic incongruities and lack of dignity in such combinations as “a constructed organization [aufgezogene Organisation].” (Klemperer, The Language of the Third Reich, 48) 12. Klemperer, The Language of the Third Reich, 159.
13. Ashwell, “Gendered Slurs”; Burnett, “A Persona-Based Semantics for Slurs”; Foster, “Busting the Ghost of Neutral Counterparts.” See also the discussion of the neutral counterpart view in Hess, “Slurs: Semantic and Pragmatic Theories of Meaning,” 455–56.
14. We do not in this volume prescribe best practices as regards when slurs might reasonably be mentioned in academic discourse. Nonetheless, our arguments in the next section, and the arguments of scholars we will cite, suggest that some slurs can retain power to harm even when quoted. For this reason, we prefer to take a cautious approach, generally avoiding mention of slurs that we feel is unnecessary for our argument, but occasionally mentioning slurs when it is helpful to use an example or when discussing an example that has been discussed by other scholars, and only doing so for slurs that, as best as we can tell, are unlikely to cause grave offense, and for which, as far as we can tell, the mention in an academic context does not in and of itself constitute to marginalization or silencing of others. Here we align with scholars Jyoti Rao (quoted in the coming discussion), and follow a policy close to that suggested by Cassie Herbert, “Exclusionary Speech and Constructions of Community,” chapter 5, “Talking about Slurs,” 130–60.
15. Bannon’s remarks were made in a speech at a rally of the Front National in France, and reported later that day in Adam Nossiter, “ ‘Let Them Call You Racists’: Bannon’s Pep Talk to National Front,” New York Times, March 10, 2018.
16. A CNN news report runs as follows: Thomas Figures, a black assistant US attorney who worked for Sessions, testified that Sessions called him “boy” on multiple occasions and joked about the Ku Klux Klan, saying that he thought Klan members were “OK, until he learned that they smoked marijuana.” (Scott Zamost, “Sessions Dogged by Old Allegations of Racism,” CNN, November 18, 2016, https://www.cnn.com/2016/11/17/politics/jeff-sessions-racismallegations) As attorney general, Sessions was active in using legal means to crack down on marijuana use, but not in using legal means to restrict discrimination.
17. Delgado, “A Tort Action for Racial Insults, Epithets, and Name Calling,” which was originally published in 1982 as “Words That Wound: A Tort Action for Racial Insults, Epithets, and Name Calling,” Harvard Civil Rights-Civil Liberties Law Review, 133–81.
18. Chideya, The Color of Our Future, 9.
19. Ice Cube (O’Shea Jackson Sr.) made his remarks on Real Time with Bill Maher, June 9, 2017.
The quote appears in Kristine Phillips, “ ‘That’s Our Word, and You Can’t Have It Back’: Ice Cube Confronts Bill Maher for Using the N-Word,” Washington Post, June 10, 2017. The quote is used to make a similar point to ours in Henderson, Klecha, and McCready, “Response to Pullum on Slurs.” 20. Patricia J. Williams, “Sensation,” The Nation, May 6, 2002, 9. Williams is objecting specifically to the extensive discussion of the N-Word in a much-cited book by Randall Kennedy that actually uses the offensive epithet as the first word of its title. Williams, after reporting on an incident in which someone brandished the book in front of her on a bus, is dubious about Kennedy’s choice of title. She comments, after describing the emotive power of the word, that “Kennedy milks this phenomenon only to ask with an entirely straight face: ‘So what’s the big deal?’ ” 21. In contrasting offense with injury, we are inspired by Charles Lawrence’s discussion of racist speech: “The word offensive is used as if we were speaking of a difference in taste, as if I should learn to be less sensitive to words that ‘offend’ me. I cannot help but believe that those people who speak of offense—those who argue that this speech must go unchecked—do not understand the great difference between offense and injury” (“Regulating Racist Speech on Campus,” 74).
22. For more extensive discussion of the role of face in slurring, see Croom, “How to Do Things with Slurs.” 23. Our analysis of the labeling function of slurs is close to that of Quill Kukla. Kukla adopts terminology from Louis Althusser and later work that includes that of Judith Butler, suggesting that slurs “hail” or “interpellate” the target, positioning them within an ideology: Slurs … are hails that, like all interpellations, recognize a subject … as having a specific identity, and thereby help constitute them as having that identity by calling upon them to recognize themselves as having it and hence as subjected to sets of norms.
Specifically, they are interpellations that recognize a subject … as having a (1) generic, (2) derogated, and (3) subordinated identity. (Kukla, “Slurs, Interpellation, and Ideology,” 19) Our suggestion that accommodation is a process of harmonization in response to external social stimuli that can be gradual, which we have linked to Klemperer’s “tiny doses of arsenic,” echoes Kukla’s idea that the interpellating impact of hails seems to assume (or presuppose) that someone has a certain identity, and that the effects of this assumption can be gradual: In hailing someone, the hail has to recognize that person as already having a certain identity, and, through what often gets called “constitutive misrecognition,” the one hailed must in fact come to be (at least incrementally more) the self she is recognized as being, by recognizing herself as properly recognized by the hail. I need not be conscious of an interpellation as an interpellation in order for it to work, but paradigmatically, upon being successfully interpellated, I have an experience of recognizing that it is “really me” who has been recognized as having the identity I have. (“Slurs, Interpellation, and Ideology,” 13) Though we follow Kukla in several respects, our analysis of slurs differs both because we set our account in terms of our own idiosyncratic approach to ideology and speech practice, based on resonance, attunement, and harmonization, but also because we consider properties of slurs that are not the primary targets of Kukla’s work, notably exigence and hyperprojectivity.
24. As Kukla puts it, “The use of a slur, whether targeted directly at its victim or used among insiders, helps generate multiple interrelated subject positions; it does not merely constitute the identity of the one slurred” (“Slurs, Interpellation, and Ideology,” 31).
25. For discussion and explanation of the reasons why there is variation in the offensiveness of different slurs, see Jeshion, “Expressivism and the Offensiveness of Slurs.” 26. McCready and Davis, “An Invocational Theory of Slurs,” 2.
27. Carroll, Alice in Wonderland, 191.
28. It is unclear to us to what extent McCready and Davis would agree with us that the linguistic mechanisms needed for an analysis of slurs are not unusual. Certainly, there is a mathematical level at which their analysis, worked out in technical detail in their paper, involves only machinery also found in the analysis of other expression types. But they would perhaps accept or even take it as a positive that many aspects of their analysis are far from being standard in formal semantics. It would take us too far afield to attempt a detailed discussion of their formal analysis here. We merely note that we see McCready and Davis’s paper as working out a somewhat parallel intuition to that driving our own arguments, and we hope that the relationship between the accounts can be better exposed in future work.
29. The importance of distinguishing communities of practice, and their associated ideologies, is important in much of the work on reclamation, for example, Hess, “Practices of Slur Use”; and PopaWyatt, “Reclamation: Taking Back Control of Words.” However, the notion of practice plays at most a minor role in most theories of the meaning of slurs in their canonical pejorative uses.
30. Nunberg, “The Social Life of Slurs,” 244.
31. Slurs derogate because they are prominent within racist communities of practice and conjure up discriminatory ideologies; this explains why intention is not essential to the power of a slur to harm. A related point is that by explaining the power of the slur in terms of a community rather than in terms of the social position or authority of a speaker, we can explain how slurs used by people lacking high social position or authority can still wound with their words, as in the case of children slurring. Our strategy of locating the power of slurs relative to a larger group than the individual speaker means that our analysis is somewhat parallel to Ishani Maitra’s analysis of hate speech (Maitra, “Subordinating Speech”). Her notion of “derived authority” allows the necessary authority for a hateful act to be licensed not by the intrinsic authority of the speaker, but by the authority the speaker inherits from a social group that they are taken to represent. Central to her analysis is the idea that this authority may effectively be given implicitly, as when onlookers to a hateful act are silent and their silence is taken as assent.
32 . For broader overviews of the slurs literature, see, for example, Popa-Wyatt, “Slurs, Pejoratives, and Hate Speech”; and Hess, “Slurs: Semantic and Pragmatic Theories of Meaning.” 33. Potts, “The Expressive Dimension”; Williamson, “Reference, Inference and the Semantics of Pejoratives.” 34. Hom, “The Semantics of Racial Epithets,” 432, ellipses in original.
35. There is much further nuance to Hom’s analysis, since, like us, he has an externalist take on meaning. Thus he allows that people using the slur might be no more able to list all the properties of those in the slur category than is someone who uses the term “beech” or elm” able to list the sufficient and necessary conditions associated with membership of different tree species (to use Putnam’s famous motivating example for externalism, once again). However, there is also a disanalogy between Hom’s presentation of his account of slurs and a Putnamian account of “elm.” Putnam did not propose writing the meaning of “elm” as a conjunction of sufficient and necessary conditions for being an elm, and we presume that if asked to write the first-order meaning of “there is an elm,” Putnam would simply have written ∃x elm(x), rather than ∃x tree(x) ∧ deciduous(x) ∧ …, although the question of how an externalist should represent a meaning depends on what the goal of representing a meaning is. To the extent that the representation is supposed to be akin to a mental representation, it seems to us that it would be odd to suppose that mental representations are similar to logical formulae with unspecified conjuncts, which Hom’s have.
If this is what the representations were for, then Hom’s representation of “Fred is a Y” (where Y is a slur) could have simply been “Y(Fred),” rather than a long and largely unspecified conjunction. Hom’s point, though, is that individual speakers can use slurs without knowing the truth-conditions of their statements, because the truth conditional meaning is sustained not by individuals alone, but by a community of speakers who share an ideology. Here, we agree with Hom.
36. To be clear, there is no clean division between philosophical and linguistic approaches to these issues, although we hope that the scholars in question would accept our broad-brushed characterizations of their work. The relevant papers are: Marques, “The Expression of Hate in Hate Speech”; García-Carpintero, “Pejoratives, Contexts and Presuppositions”; Marques and GarcíaCarpintero, “Really Expressive Presuppositions”; Schlenker, “Expressive Presuppositions”; Cepollaro, “In Defence of a Presuppositional Account of Slurs.” Related to the latter is Cepollaro and Stojanovic, “Hybrid Evaluatives,” which generalizes the presuppositional approach to slurs to socalled thick terms, i.e., broadly the sorts of evaluative adjectives we discuss in section 7.3.
37. Other examples of accounts in which the semantics of slurs is identified with a neutral counterpart and pragmatic mechanisms are evoked to explain further properties of slurs are Nunberg, “The Social Life of Slurs,” and Bollinger, “The Pragmatics of Slurs.” Diana Blakemore offers a pragmatic variant in which the offensiveness of slurs is derived from “meta-linguistic knowledge that the word is an offensive means of predicating and referring” (“Slurs and Expletives,” 34). She is explicit about the semantic content of slurs being identical to that of a neutral counterpart, for example stating that the words “dyke” and “lesbian” have identical semantics, these being exactly the cases that, as we will discuss, Heather Burnett convincingly argues cannot be semantically identical. 38. Hedger, “The Semantics of Racial Slurs”; Jeshion, “Expressivism and the Offensiveness of Slurs”; Hess, “Slurs and Expressive Commitments.” Note that although Jeshion is committed to slurs having a neutral semantic core, her position on whether there are neutral counterparts is nuanced, first because she thinks that in slurring the extension of a term is sometimes contracted and sometimes extended, and second because she suggests that at least some slurs lack clear counterparts.
Although Kent Bach (“Loaded Words: On the Semantics and Pragmatics of Slurs”) contrasts his account of slurs with expressivist accounts, terming his view “loaded descriptivism,” his view is related, and he explicitly makes the same assumption that slurs share a neutral core semantics with a neutral counterpart. Yet another account that explicitly separates the semantic component of slurs from an expressive component is that of Richards, When Truth Gives Out, especially chapter 1, “Epithets and Attitudes,” 12–41. Richards offers an account in which slurs have both conventionalized expressive and conventionalized performative components in addition to a semantic component. He does assume there are neutral counterparts, noting, for example, “It is just not open to me to unilaterally detach the affect, hatred, and negative connotations tied to most slurs and use them interchangeably with their neutral counterparts” (41). However, Richards does not assume that the semantic value of slurs is identical to that of the neutral counterpart; rather he denies that sentences containing slur terms are even truth-evaluable. 39. Hess, for example, writes, “The derogatory meaning is not part of the semantic content of a slur (which is identical to its neutral counterpart)” (“Slurs and Expressive Commitments,” 280).
40. Burnett, “A Persona-Based Semantics for Slurs.” 41. Peter Gärdenfors, The Geometry of Meaning. As briefly noted in chapter 3, there is an abstract similarity between his model and ours, both using spatial metaphors. Gärdenfors’s account of mental representation is far more technically elaborated than our own, though he does not attempt to model what we term dispositional attunements, and he does not apply his model in the political or social sphere.
42. The term hyperprojective, used as a description of the special tendency of slurs to have effects even when quoted, is taken from Elizabeth Camp (“A Dual Act Analysis of Slurs”).
She says that a perspective (which functions in her theory much like ideologies and perspectives do in ours, with affective as well as cognitive dimensions) is “hyper-projective” in that “it … typically projects across indirect attitude and speech reports, which are supposed to be projection ‘plugs’ ” (39).
43. Wittgenstein, Philosophical Investigations, 8, proposition 18.
44. Tirrell, “Discursive Epidemiology,” 117. While we take our approach to be compatible with Tirrell’s, her focus is on the progressive harm caused by repeated encounters with toxic speech (an important topic for us too), and not so much on the mechanism by which an individual act produces harm, a focus in the current section. As we understand her account, the extreme negative effects that an individual speech act can have should be understood as resulting from an intolerable buildup of toxic stress, like the proverbial straw breaking a camel’s back. This way of thinking usefully adds to the account we give here, in terms of attention, presupposed ideology, and emotional attunement.
45. Matsuda, “Public Response to Racist Speech,” 2336–37. The Patricia Williams reference is to Williams, “Spirit-Murdering the Messenger.” 46. Lawrence, “Regulating Racist Speech on Campus,” 70.
47 . Popa-Wyatt and Wyatt, “Slurs, Roles and Power,” 2897.
48. Rao, “Observations on Use of the N-word in Psychoanalytic Conferences,” 317–18. 49. Jeshion, “Expressivism and the Offensiveness of Slurs”; Richard, When Truth Gives Out (especially chapter 1, 12–42).
50. Popa-Wyatt and Wyatt, “Slurs, Roles and Power,” 2897.
51. Rao, “Observations on Use of the N-word in Psychoanalytic Conferences,” 316–17.
52. Butler, Excitable Speech, 100.
53. Cepollaro and Zeman, “Editors’ Introduction: The Challenge from Non-Derogatory Uses of Slurs,” 1. Note that for Cepollaro and Zeman, as for Camp, the phenomenon of hyperprojectivity includes both projectivity from direct quotations, but also from indirect quotations, for example of speech reports. Thus it is consistent (although, as we have argued, empirically inadequate) to claim that some effects of slurs are hyperprojective while still maintaining that quotation marks block projection. Note also that the nonderogatory uses of slurs considered in the special issue include reclaimed uses of slurs, which we discuss briefly below, and which we agree are nonderogatory.
54. The extent to which the effects of slurs project from quotation is an empirical issue, so one might look to work in the burgeoning field of experimental semantics and pragmatics for evidence.
However, slurs are problematic to work with experimentally, since they might cause harm to experimental subjects, and perhaps for this reason there is a paucity of experimental evidence. There are, to our knowledge, no experimental studies within linguistic semantics and pragmatics directly testing the feelings evoked in the targets of real-world slurs (as opposed to artificial pseudo-slurs), and that is perhaps as it should be. However, there is at least one study that looks at the perceived offensiveness of slurs in direct and indirect quotations, albeit that experiment does not control for whether experimental subjects were in the target group of the slur: Cepollaro et al., “How Bad Is It to Report a Slur? An Empirical Investigation.” As we read the results reported in this paper (38), they show (i) that direct quotes of others using various Italian pejorative expressions are approximately as offensive as unquoted occurrences of the same expressions, (ii) that indirect quotations involving the expressions are almost as offensive, and (iii) that both are far more offensive than nonpejorative expressions. However, the authors do not draw any conclusions from the fact that stimuli involving quotations of slurs, in forms like “Y: ‘X is a P’ ” (37), are seen as similarly offensive to slurs that were not within quotation marks, which presumably relates to what they took subjects to be judging (what was quoted vs. the presentation of the quote).
As regards indirect quotations, the authors conclude that “utterances featuring slurs or non-slurring insults are perceived as less offensive when they occur in indirect reports, even though the report cannot entirely delete the offensiveness.” This formulation suggests that the offensiveness of slurs is very much reduced in indirect quotations, but that it is in tension with the data they report on average levels of offensiveness. If we were to assume for simplicity that the offensiveness scale is linear, and take their mean ratings for “non-slurring labels” as a baseline, we find that indirect reports of slurs are over 70 percent as offensive as direct reports. While it is unclear what numerical conclusions are warranted, it seems to us that a faithful description of the results should perhaps not be that reports “cannot entirely delete the offensiveness,” but rather that reports of slurs only slightly diminish the offensiveness.
55. According to most accounts, presuppositions are conditions that should be met, and are thus propositional. Manuel García-Carpintero and Teresa Marques allow that presuppositions can also be affective. Since we have presented in this book a model in which presuppositional resonances have just this property, we obviously agree. García-Carpintero and Marques take the affective nature of slur presuppositions to explain hyperprojectivity, but it is not clear to us exactly how the affective nature of slur presuppositions explains differences in projectivity. Certainly, if the special projectivity of affective presuppositions were to be stipulated, rather than independently explained, then this approach too would amount to special pleading. See Marques, “The Expression of Hate in Hate Speech”; García-Carpintero, “Pejoratives, Contexts and Presuppositions”; Marques and GarcíaCarpintero, “Really Expressive Presuppositions.” For a related view, with a similarly structured explanation of hyperprojectivity effects, see also Schlenker, “Expressive Presuppositions.” 56. Cepollaro, “In Defence of a Presuppositional Account of Slurs.” 57. For experimental data confirming the generally highly projective nature of conventional implicatures, but also the more general variability of projection effects, see Tonhauser, Beaver, and Degen, “Gradience in Projectivity and At-Issueness.” 58. “Democratic Presidential Candidates Debate at Drexel University in Philadelphia, Pennsylvania,” October 30, 2007, retrieved February18, 2023, The American Presidency Project, https://www.presidency.ucsb.edu/documents/democratic-presidential-candidates-debate-dre xeluniversity-philadelphia-pennsylvania.
59. Matsuda et al., Critical Race Theory, 8.
60. Butler, Excitable Speech, 36.
61. Swanson, “Slurs and Ideologies,” 9–14 62. Jeshion, “Slurs, Dehumanization, and the Expression of Contempt,” 84.
63. In Down Girl, Kate Manne argues that women who behave according to patriarchal norms are rewarded in the attitudes taken toward them.
64. Jeshion, “Expressivism and the Offensiveness of Slurs,” 321.
65. Jeshion, “Expressivism and the Offensiveness of Slurs,” 322. Note that the difference in strength of the two slurs Jeshion was considering in this passage follows from our definition of slur strength, above.
66. One of Jeshion’s arguments (in “Expressivism and the Offensiveness of Slurs”) against the conventionality of the link between uses of these slurs and their associated histories of racial domination is that their neutral counterparts, “Black” and “white,” when employed with a contemptuous tone, can also have the same perlocutionary effect. We agree with the datum—but the contemptuous tone is a conventional signal that the supposedly neutral word is being used to exemplify a speech practice that is part of the “slurring” ideology. The fact that there is a prominent but sometimes contested practice of capitalizing the first letter of one of these terms but not the other makes it obvious that ideology is tied to the use of either of them: capitalization suggests that the adjective is not merely denoting a property of certain individuals, but marking out a sociopolitically significant classification, by analogy with the capitalization of nationality adjectives and others derived from proper names.
67. Anderson and Lepore, “Slurring Words”; “What Did You Call Me?”; Lepore and Stone, “Pejorative Tone”; Anderson, “Philosophical Investigations of the Taboo of Insult.” 68.
Munton, “Prejudice as the Misattribution of Salience.” Her invocation of attention has been inspirational for us, not only in centering attention in our analysis of slurs, but also as regards the model of harmony in terms of landscapes of attunements in chapter 3. Munton’s attentional structures play a similar role to the structuring of attunements in our model, determining when activation of one attunement will tend to lead to activation of another.
69. Thanks to Endure McTier for discussion of “Karen” as a slur.
70. Anderson, “Calling, Addressing, and Appropriation.” Further discussion of reappropriation is found in Bianchi, “Slurs and Appropriation”; Jeshion, “Pride and Prejudiced”; Popa-Wyatt, “Reclamation: Taking Back Control of Words”; Quaranto, “How to Win Words and Influence Meanings”; and Ritchie, “Social Identity, Indexicality, and the Appropriation of Slurs.” 71. Famously, in Excitable Speech, Butler urges that the reclamation of hurtful speech in this way is the best way forward in the face of hate speech, most clearly in the section “Hate Speech/State Speech” (96–102) where she says, for example, “The possibility of decontextualizing and recontextualizing such terms through radical acts of public misappropriation constitutes the basis of an ironic hopefulness that the conventional relation between word and wound might become tenuous and even broken over time” (100).
72. Dembroff, “Beyond Binary.” 73. “Representative Alexandria Ocasio-Cortez Addresses House on Exchange with Representative Ted Yoho,” C-SPAN, July 23, 2020.
74. McGowan, Just Words, 110–11.
75. Swanson, “Slurs and Ideologies,” 1–2.
76. Dangerous Speech Project, Dangerous Speech: A Practical Guide, 2022, https://dangerousspeech.org/guide/.
77. Tirrell, “Genocidal Language Games,” 192.
78. The effect of calling a New York Times columnist turns out to depend substantially on how the target of the insult responds. When in August 2019 an academic, David Karpf, called New York Times columnist Bret Stevens a “bedbug” on Twitter, there was little initial reaction. However, after Stevens sent an email to Karpf (cc-ed to Karpf’s provost) suggesting that Karpf call him “bedbug” to his face, and Karpf copied the email in a tweet, the exchange suddenly drew largescale attention, largely at Stevens’s expense, leading Stevens to quit Twitter (Luke O’Neil, “NYT Columnist Quits Twitter after Daring Critic to ‘Call Me a Bedbug to My Face,’ ” Guardian, August 27, 2019). Despite the story having brief notoriety, we see no evidence of long-term take-up of the practice of labeling newspaper columnists or other center-right political commentators as bedbugs.
79. Muhammad, The Condemnation of Blackness, chapter 2, “Writing Crime into Race: Racial Criminalization and the Dawn of Jim Crow,” 35–87.
80. Timofei Sergeitsev, “What Should Russia Do with Ukraine?,” RIA Novosti, April 5, 2022.
See also Mariia Kravchenko, “What Should Russia Do with Ukraine? [translation of a propaganda article by a Russian publication],” Medium, August 24, 2022, https://medium.com/@kravchenko_mm/what-should-russia-do-with-ukraine-translation-of-apropaganda-article-by-a-russian-journalista3e92e3cb64.
81. Snyder, “Russia’s Genocide Handbook,” news page of McGrublian Center for Human Rights, Claremont McKenna, retrieved February 25, 2023, https://human-rights.cmc.edu/2022/04/14/russiasgenocide-handbook/.
82. Goebbels, “Communism with the Mask Off (1935),” 128.
83. See Snyder, The Road to Unfreedom, especially chapter 4, “Novelty or Eternity,” 111–58.
84. Our discussion in the rest of this section overlaps and borrows from Jason Stanley and Federico Finchelstein, “White Replacement Theory Is Fascism’s New Name,” Los Angeles Times, March 24, 2022, https://www.latimes.com/opinion/story/2022-05-24/white-replacement-theoryfascism-europehistory. We are grateful to Finchelstein for the Hitler and Mussolini references.
85. Hitler, Mein Kampf, 826–27.
86. Benito Mussolini, “La Razza Bianca Muore?,” La Stampa, September 5, 1934, retrieved February 25, 2023, https://ia601805.us.archive.org/24/items/lastampa_1934-09-05/lastampa_1934- 09-05.pdf. 87. Jason Stanley and Fredericho Finchelstein, “White Replacement Theory Is Fascism’s New Name,” Los Angeles Times, May 24, 2022, https://www.latimes.com/opinion/story/2022-05- 24/white-replacement-theory-fascism-europe-history.
88. Grant, The Passing of the Great Race. He writes, “The man of the old stock is being crowded out of many country districts by these foreigners, just as he is to-day being literally driven off the streets of New York City by the swarms of Polish Jews” (81).
89. The missing third spirit in the French trinity of democracy is of course fraternité, a value that tends to be espoused more at the ends of the political spectrum than in the center.
“Brotherhood” suggests a fusion of identity whereby harm to one is harm to all, hence “brothers in arms” and its attendant military resonances. On the left, “brother” serves as a variant of “comrade,” while the history of the far right is littered with violent political groups labeled, with undisguised sexist intent, “Brotherhood of.…” 90. Grant, The Passing of the Great Race, 3–10.
91. David French, “Structural Racism Isn’t Wokeness, It’s Reality,” Dispatch, July 25, 2021, https://frenchpress.thedispatch.com/p/structural-racism-isnt-wokeness-its.
92. Corredor (“Unpacking ‘Gender Ideology’ and the Global Right’s Antigender Countermovement”) gives an excellent description of the history of the propaganda campaign, theorized as a countermovement.
93. From Morrison’s Nobel Prize address, Stockholm, December 7, 1993.
94. We discuss here bureaucratic speech as a state language that is used to control a national population, but we do not mean to suggest that it is the only type of language that fulfills such functions politically. Neither do we mean to imply that those seeking to maintain control in different nation-states and political regions uniformly use the same rhetorical methods. On the contrary, we suppose that the most effective language of control in a given sphere depends heavily on cultural resonance, varying according to local history, social organization, and ways of seeing and talking about the world. Here we think of the work of Lisa Wedeen, who sets detailed ethnographic studies of individual-level discourse in Syria and Yemen in the context of national political situations. An example of a type of discourse we don’t discuss in detail in this volume, but that she shows is integral to political discourse in Assad’s Syria, is discourse that extends the metaphor of family to a national and political level (Ambiguities of Domination, 49–65). Authoritarian regimes commonly use familial metaphors and familial framings, in particular the idea of the leader as father to the nation, but the use of family in Syria is extended: “The official narrative communicates understandings of obedience and community in terms of a chain of filial piety and paternal authority that culminates, and stops, in Asad.” Further, while the “invocation of family relationships may be more or less resonant to individual Syrians,” it is nonetheless the case that Syrians perforce must partake in the discourse every day: “To be ‘Syrian’ means, in part, to operate within this rhetorical universe” (65).
95. Alexopoulos, Illness and Inhumanity in Stalin’s Gulag, 208.
96. Alexopoulos, Illness and Inhumanity in Stalin’s Gulag, 210–11.
97. Alexopoulos (Illness and Inhumanity in Stalin’s Gulag, 241) writes, “The MVD-Gulag leadership wanted camps to report around 70–80 percent of prisoners in Group A working in industry, so camps were sure to classify this many as ‘basically fit for physical labor.’ ” 98.
Alexopoulos, Illness and Inhumanity in Stalin’s Gulag, 86.
99. Dean Kirby, “Putin Sends Mariupol Survivors to Remote Corners of Russia as Investigation Reveals Network of Camps,” inews, May 7, 2022, https://inews.co.uk/news/putin-mariupolsurvivors-remote-corners-russia-investigation-networ k-camps-1615516.
100. For references and extended discussion of the role of emergency managers in the Flint water crisis, see Stanley, “The Emergency Manager: Strategic Racism, Technocracy, and the Poisoning of Flint’s Children.” 101. Curt Guyette, “The EAA Exposed: An Investigative Report,” Detroit Metro Times, September 24, 2014, http://www.metrotimes.com/detroit/the-eaa-exposed-aninvestigativereport/Content?oid=2249513.
102. Marlin, Propaganda and the Ethics of Persuasion, 38. Ellul himself writes of the effects of bombarding the public with facts and figures: “What remains with the individual affected by this propaganda is a perfectly irrational picture, a purely emotional feeling, a myth. The facts, the data, the reasoning—all are forgotten, and only the impression remains.… Thus propaganda in itself becomes honest, strict, exact, but its effect remains irrational” (Propaganda: The Formation of Men’s Attitudes, 86).
103. Thanks to Scott Shapiro for calling attention to this on Twitter: @scottjshapiro, Tweet, May 8, 2022, https://twitter.com/scottjshapiro/status/1523309136540098560.
104. Alexopoulos, Illness and Inhumanity in Stalin’s Gulag, 4.
105. Collins, Black Feminist Thought, 255.
106. Arendt, Eichmann in Jerusalem: A Report on the Banality of Evil.
CHAPTER ELEVEN Free Speech The First Amendment was conceived by white men from the point of view of their social position. Some of them owned slaves; most of them owned women. They wrote it to guarantee their freedom to keep something they felt at risk of losing. Namely—and this gets to my next point—speech which they did not want lost through state action.
They wrote the First Amendment so their speech would not be threatened by this powerful instrument they were creating, the federal government. You recall that it reads, “Congress shall make no law abridging … the freedom of speech.” They were creating that body. They were worried that it would abridge something they did have. You can tell that they had speech, because what they said was written down: it became a document, it has been interpreted, it is the law of the state.
—CATHARINE MACKINNON1 RUSSIA’S 2013 GAY PROPAGANDA LAW banned the promotion of nontraditional lifestyles to minors. This law had a devastating effect on Russia’s society, sending its LGBTQ citizens deeply into the closet for self-protection, and preventing Russian citizens generally from understanding LGBTQ perspectives. Robust free-speech protections would prevent such a law.
There is scarcely any question that such principles are vital in protecting freedom. Banning books, restricting history, and suppressing the teaching of minority perspectives—all of these acts are inimical to a free society. And yet, free societies are often imperiled by propaganda, which is allowed by robust free-speech protections. Let’s call this the paradox of democracy.
In this chapter, we use our model to sharpen this paradox.
In the bulk of this chapter, we canvass a number of arguments for freespeech principles of various scopes. In each case, we will show that the arguments make problematic idealizations. We conclude by tentatively endorsing a familiar suggestion for refocusing debate about the regulation of speech. There are a variety of justifications for freedom of speech. These arguments have different scopes, in the sense that they are set out to protect different zones for a free-speech principle. For example, Alexander Meiklejohn, emphasizing the point that “the First Amendment … is not the guardian of unrestricted talkativeness,” aims to protect speech providing relevant viewpoints on topics that citizens are deciding upon in joint deliberation of policy, as if in a town meeting.2 In what follows, we consider various justifications for free-speech principles in the light of our theoretical framework.
Justifications of free speech also differ in the basic moral frameworks upon which they draw.
The most salient division is the one between consequentialism and deontology. The largest class of argument for free speech takes the form of a consequentialist argument—having a free-speech ideal will result in society having, collectively, more of a desirable property: knowledge, for instance, or democratically functioning institutions. Other arguments are based not on effects, but on the reasonableness of speech in itself. This is the case for deontological arguments based on autonomy, conceived as something to which we have a right—a right is something we have regardless of its consequences. A consequentialist argument appeals to very different kinds of premises than the autonomy argument. You won’t, for example, find premises about rights in consequentialist arguments. And you won’t find premises about what maximizes some good property in deontological arguments. If we restrict ourselves either just to consequentialist or autonomy arguments for free speech, we find they too differ in the attendant commitments of their premises. Despite this variety, we will argue that all of them make false idealizing presuppositions about speech.
The view that speech can be inimical to democracy has a long history. In book VIII of Plato’s Republic, Socrates argues that democracy is the form of government most likely to lead immediately to tyranny. The argument that democracy leads directly to tyranny appeals to the ideal of free speech, which Plato rightly regards as central to the system of democracy.
According to Plato, democracy leads to tyranny, because the character of its citizens and its central values enable the rise of a tyrant and virtually guarantee their success: [The people’s protector’s] docile followers grant him so much free rein that he shrinks from nothing, even shedding the blood of his own kin. He resorts to the customary unjust accusations in order to haul the victim into court and then to take his life. So he murders, and with impious tongue and lips consumes his own flesh and blood.3 Democracy’s liberties enable all character types to vie for power. Using free speech, the would-be tyrant lies and spreads fear, presenting himself as “the people’s protector” against a supposedly dangerous group. A tyrant has no barrier against using dangerous propaganda to seize power. Democracy’s liberties allow a tyrant to seek this path, using speech that harms democracy.
Plato’s concern here is with propaganda and its role in creating what we have termed an antagonistic ideological social group. The natural home of political propaganda is in the drum up to war. The tyrant must ceaselessly promote the people’s sense that they “will always be in need of a leader” and thus “undertake[s] an ongoing search for pretexts to make war.”4 Plato is arguing that to stay in power, the tyrant must use language to foment ungrounded fear of others. A system based on liberty as a political ideal must allow those most unsuited to be political leaders to seek that role. And a system based on liberty as a political ideal allows them to use demagogic propaganda to seize it. Plato’s argument against liberty as a political ideal depends upon strong assumptions about the power of speech to successfully instill democratically problematic emotions.
Plato is right to warn of speech that threatens freedom. The existence and ubiquity of such speech poses a serious threat to arguments for freespeech principles. 11.1. On the Benefits of Free Speech The classical philosophical defense of free speech is “Of the Liberty of Thought and Discussion,” chapter 2 of John Stuart Mill’s On Liberty. The guiding principle of Mill’s inquiry is the Harm Principle laid out in his first chapter: The only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others.5 In chapter 2, Mill offers a consequentialist argument, the goal of achieving the common good being his primary justification in support of free speech.
A society with the widest possible freedom of speech will be one that ends up having both better justification for its guiding principles as well as citizens who are more knowledgeable and more able to form correct judgments about policy. These goods must be weighed, of course, against potential harms of having free speech in Mill’s broad sense.
Chapter 2 of On Liberty concludes with brief consideration of a “more fundamental objection” to freedom of opinion.6 The objection involves the damage of “intemperate discussion,” including “invective, sarcasm, personality.” Mill regards these damages as less serious in discussion between equals; for Mill, “whatever mischief arises from their use is greatest when they are employed against the comparatively defenseless.” Contrarily, Mill traces the harm (“mischief”) of invective and sarcasm to differential power relations—when these tools are directed against the defenseless.
For example, imagine an entrenched religious orthodoxy that has seized the power in the institutions and directs invective against outsiders who raise skepticism against it, or wealthy aristocrats ridiculing impoverished workers who seek better working conditions. A power relation of the one class over the other makes speech harmful, in ways that make the situation harder, rather than easier, for improving the society democratically or otherwise. In short, a concern for Mill’s argument for free speech (here, that it does not violate his “harm principle”), is the existence of power differentials in a society. He believes such differentials allow speech to be worrisomely harmful, and in such a way as to impact directly the benefits of free speech.
This suggests that Mill regards authority as necessary to explain harmful speech in these cases. According to Mill, speech is given the power to harm because of large differences in power between people. There is now a large literature devoted to evaluating the question of whether the harm of speech depends on power relations in this way. The topic of authority and subordinating power has been much discussed in more recent philosophy. Rae Langton’s “Pornography’s Authority? Response to Leslie Green” begins “Subordinating speech is authoritative speech,” continuing a theme from her classic 1993 paper, “Speech Acts and Unspeakable Acts.”7 Using speech act concepts from J. L.
Austin, Langton argues that a vital “felicity condition” for subordinating speech is a speaker’s authority. Speech is harmful because of the power relations of one group over another.
Langton and Mill are no doubt correct that authority can make subordinating speech more harmful. Many of the legal cases involving harmful speech manifest power differentials of the kind that threatens the democratic benefits of speech, for example, employers leveling harmful racist speech at employees who promote adherence to union rules.8 Other legal cases involve White employees accosting fellow Black employees with racist slurs. However, it is far from obvious that authority of the speaker is necessary for speech to be genuinely harmful.
In Charles Lawrence’s paper in the seminal critical race theory volume from 1993, Words That Wound, he discusses an incident that affected his family in their home of Wilmington, Delaware. In the school that his nephews attended, four of their peers had drawn racist and antisemitic slogans, pictures of hooded Klansmen, Nazi swastikas, and a burning cross.9 The slogans and symbols had been drawn by teenagers, with presumably little authority. But Lawrence makes clear that the speech was harmful nonetheless, even on those in greater positions of authority. And there are many such discussions of harmful speech without apparent authority in the critical race theory literature. In Mari Matsuda, Charles R.
Lawrence III, Richard Delgado, and Kimberlé Williams Crenshaw’s introduction to Words That Wound, the case of harmful speech they begin by discussing is the Ujamaa incident at Stanford University, in which two White freshmen defaced a Beethoven poster, creating a caricature of a Black man, with racist slurs scrawled across it. This is also not obviously an example in which authority, at least not officially sanctioned authority, is central to the harm.10 More recently, Ishani Maitra has argued against the authority claim, on the grounds that a homeless person on a subway with little social authority can yell racist and Islamophobic slogans and still engage in subordinating speech.11 The case is analogous to one we have considered (section 5.2 and section 10.2), that of a child using a slur that they don’t fully understand.
We argued that a slur can retain the power to wound independently of the understanding or intention of the speaker, because the ability of the slur to draw attention to a despised categorization rests on the discriminatory ideology of the community of practice within which the slur is used, and so does not depend crucially on the speaker’s mental state. In the modern world, we can think of social media, where many random Twitter accounts, none particularly high in authority, combine to create a subordinating speech environment against their target. Again, the mental state of individuals who tweet is not crucial. Indeed, the question of whether some of those producing tweets even have mental states is a vexed one, since they may be bots.
One can easily maintain the view that an imbalance in power relations is necessary for speech to harm, in the face of these examples. It’s plausible that power relations are evoked not by the individuals—teenagers drawing on walls, or a homeless person on a subway—but by the ideologies their actions invoke. It is because Black Americans are oppressed that drawing pictures of hooded Klansmen or hanging nooses from trees harms—the White supremacist ideology is powerful. Powerless agents can invoke in their actions and words powerful ideologies.12 As we suggested in section 10.2, “an important characteristic of the resonance of a slur is that it replicates power relationships,” and that characteristic extends to other symbols of discriminatory ideology, rather transparently in the case of the representation of members of a militia or of a standardized mode of oppressive violence.
Mill’s concern is not ultimately with the source of harmful speech, but the fact of harmful speech. Mill regards the power of speech to harm so seriously that he thinks the consequentialist arguments provided in the chapter do not suffice to override them. Did Mill provide a compelling solution to this “more fundamental objection” to the freedom of speech?
Mill’s solution to the problem he raises is a “real morality of public discussion.”13 What Mill meant here was that in order for a broad system of free-speech laws to be in place, there must be a system of norms that prevents people in authority from using their power and status to dominate others in conversation. In short, Mill’s argument for free speech presupposes what we have called a dialogically healthy environment, one in which joint negotiations of meaning are possible. Perhaps a system of morality of public discussion was robust among the gentlemen in Victorian England in his day—we cannot adjudicate the matter here. But we live in an emotionally, epistemically, and behaviorally polarized society.
That is, we live in a society that is not a dialogically healthy environment. It’s therefore fair to wonder whether his arguments have contemporary relevance today in our various conflicts about the extent of free-speech principles.
In the presence of a “real morality of public discussion,” Mill argues that societies guided by a free-speech ideal are likely to maximize epistemic benefits, without violating the harm principle. Mill’s argument depends on robust and controversial claims about extralegal but socially powerful systems of norms. These kinds of systems of norms are manifestly lacking today.
Robert Post has an influential account of how to preserve the epistemic benefits of free speech in the presence of an information space filled with conspiracy theories and other malign speech that threatens the public information space.14 Post argues for the importance of scientific institutions and universities, maintaining that intelligent self-government requires experts to be shielded from state control, because the democratic competence of participants in public discourse requires their having “access to disciplinary knowledge.”15 Post’s thought is that scientific research institutions, by conditioning discourse to rules of scientific debate, when wedded to a raucous public sphere, can lend order and rigor to channel the chaos of open debate in ways that allow the truth to win out. Is this a plausible response to the epistemic threat of oppressive speech?
By securing access to expert knowledge, and thereby securing “democratic competence” of citizens, one might hope to ameliorate the harms of oppressive speech.16 We are sympathetic to this view, but it creates a tension. As we have seen, one kind of oppressive speech is bureaucratic speech. When bureaucratic speech oppresses, it does so by masking harmful ideologies and realities. In its positivist version, it is in tension with democracy, since it involves active resistance to accommodating nonexperts.
In Frederick Hoffman’s 1896 book, Race Traits and Tendencies of the American Negro, he argues that Black people have less “vital force” than White people.17 This is a work of scientific racism, filled with statistics and massive amounts of evidence. Hoffman devotes an entire chapter largely to showing that Black people have “excessive mortality.”18 In another, he argues that Black people have vastly greater propensity toward criminality.
In each case, he claims that there is no environmental explanation—for example, he argues that “in Washington, the colored race has had exceptional educational, religious, and social opportunities,” in order to show that poor environment cannot explain a lack of “moral progress” or racial differences in arrests.19 In his discussion of mortality, he argues that the relevant White and Black populations in his studies have the same environmental conditions.
Hoffman’s book is presented as the epitome of scientific reason. But these ideals of reason are hijacked by racist ideology. As the historian Khalil Muhammad argues, the vocabulary of crime and disease resonated with anti-Black racist ideology during the end of the nineteenth century (as it does today).20 The use of this and related vocabulary in this context attuned people to racist ideology, thereby masking social and environmental causes of crime and disease. When Hoffman turns to the topic of White European immigrant populations and their high crime, disease, and divorce rates, these environmental factors suddenly and vividly appear to him as obvious explanations.
In his work, Hoffman argues that White European immigrant populations in Northern cities face terrible social and environmental conditions, and that this explains their high crime rates. What is happening?
Recall Jessie Munton’s definition of prejudice as a structure that gives undue salience to certain properties (discussed in section 7.6).21 When crime by Black people is at issue, Hoffman (and anyone with anti-Black racist attitudes) gives their race undue salience. To a racist, even an unconscious one, in the case of crime by Black people, race “leaps out” as an explanation. This has the effect of masking the environment as a cause of crime. But when the topic is high crime rates in White immigrant populations, Hoffman is suddenly able to recognize environmental conditions that explain social problems. Stephen Jay Gould provides similar explanations for other examples of scientific racism. For example, in discussing Samuel George Morton’s craniometry studies, he shows that Morton gathered his evidence on the assumption that Black skulls would be smaller, tossing aside large Black skulls as corruptions of the data pool.22 The proposal we have been considering is that experts’ knowledge, if protected from state control and accessible to citizens, will filter out problematic ideologies that impede rationality. But technocratic and academic discourse can carry along these ideologies, serving as the vehicle of masking.
The long history of scientific racism makes it clear that experts have used their association with institutions of scientific research not to cut through the cacophony of public debate, but to advance biased arguments, whose bias is concealed by scientific vocabulary that is vibrant with resonances that unconsciously affect the scientific community (as in the language of crime, or IQ). As Gould emphasizes, just as in our discussion of bureaucratic speech in the previous chapter, the scientific background of the scientists he discusses—leading figures of their time—served not to pierce racist ideology, but to mask it.
Bureaucratic speech is, as we have discussed at length, a particularly powerful vehicle for justifying harmful ideologies.
Rather than exposing harmful ideologies, it often conceals them.
In Epistemic Injustice, Miranda Fricker defines the central case of what she calls “testimonial injustice” as a credibility excess or deficit due to a social identity or class difference—with characteristic examples being deficits in the credibility given to women due to prejudices about gender, or to people because of their race or class.23 If a society has traditionally subjugated a group, preventing them from accessing the higher reaches of formal education and limiting the opportunities for the few who make their way through the substantial initial obstacles, equating expertise with institutional power would lead to epistemic injustice against this group, whose lack of access to the elevated status of expert would be read into the flawed stereotype of their group as worthy of less credibility.
In the Black feminist tradition, we find a clear warning about the dangers of giving additional authority in the form of “democratic competence” to experts. Patricia Hill Collins argues that appeal to the special authority of scientific expertise, what she calls “Eurocentric knowledge validation processes,” has been used to discount the valid testimony of Black women speaking from their own experiences, who have been denied educational opportunities.24 Following Collins, Kristie Dotson has singled out a distinctive species of testimonial injustice related to misuses of democratic competence. Contributory injustice, in Dotson’s terms, occurs when an agent’s testimony is discounted or reduced in credence solely because it is not backed up with the proper “official” epistemological resources—in this case, the ones provided by data and statistics.25 This does not mean, of course, that data and statistics are “White.” But restricting public reason to this vocabulary excludes people from discussion who are without access to educational institutions because of structural oppression. As we know, for example, from recent US discussions of the history of policing, this has been a terrible error. As the historian Elizabeth Hinton shows, poor Black communities in the late 1960s living in urban housing projects were vividly clear about the problematic and racist nature of policing practices in their communities, and they suggested changes to policing practices that are now widely regarded as important.26 Their suggestions were not heeded or taken seriously; they were discredited because of their source.
As we argued in the previous chapter, bureaucratic speech creates a community of practice that is structured by active resistance to accommodating the views of those outside the circle of expertise. In considering the relationship between democratic competence and expert knowledge, we shouldn’t abstract away from the kind of bureaucratic speech that serves as a barrier to the deliberative environment that is central to democracy.
There are clear cases in which we do want to rely on scientific expertise in policy making, where oppressive speech is less of an issue—the science of climate change or vaccines are examples. But the science that bears most directly on politics—social science—certainly can and often has functioned to mask harmful ideologies. Social scientific discussions of crime have proven at times shockingly prone to racist bias, suggesting, as Khalil Muhammad argues, that the language of crime resonates with Blackness.27 In 1996, exactly one hundred years after the publication of Hoffman’s influential work predicting the imminent demise of the United States’ Black population, the social scientist John DiIulio advanced his influential “SuperPredator Theory,” which maintained the United States faced a coming wave of hardened youth criminals, “as many as half” of whom could be “young Black males.” DiIulio predicted a massive wave of violent crime in the United States from 1996 to 2010.28 In fact, by 2010, US violent crimes rates, which had steadily and rapidly decreased since the early 1990s, were at historic lows. In an article thick with statistics, the Princeton professor made bold and confident predictions about a wave in crime despite the fact that crime had been rapidly declining for at least three years, a fact noted at the very beginning of his 1996 article. Essentially, DiIulio’s prediction of a crime wave was due to a rising demographic of Black youth—and DiIulio’s assumptions that crime rates in populations do not vary with changing socioeconomic circumstances. The bizarre nature of this assumption, embedded in an article that begins by acknowledging that violent crime had been decreasing for several years, was masked by racist resonances of the vocabulary of crime.
The racist resonances of the vocabulary of crime are measurable. We have defined something to be a resonance of a practice to the extent that it has increased likelihood of occurring in the context of that practice. If one looks at occurrences of the crime-related words “crime,” “criminal,” and “gang member” in books published in the United States, one finds that they are all three to five times as likely to be preceded by “[B/b]lack” than by “[W/w]hite.”29 By comparison, terms associated with non-Black ethnicities or races occur at zero or entirely negligible rates in such contexts. In the context of the word “rapist,” the word “[B/b]lack” is nearly ten times as likely to occur as the word “[W/w]hite,” and no other word for a race or ethnicity occurs prominently in this context. There is a notable special case of the plural “rapists.” From 2015, the frequency of the word “Mexican” suddenly shot up, so that it came to parallel the frequency of “[B/b]lack” before “rapists,” with both occurring at well over double “[W/w]hite,” and with no other ethnic, racial, or national group being prominent. That spike was created single-handedly by Donald Trump, who in the speech declaring his candidacy for US President in 2015 said of Mexican immigrants “They’re rapists” (with the caveat, “And some, I assume, are good people”).
Most of the later usages of “Mexican rapists” we see are traceable to Trump’s openly racist presentations of a violent stereotype of Mexican immigrants in the United States.
Let us make explicit a comment on this data that otherwise might be uncharitably taken out of context: the changing rate of occurrence of racial, ethnic, and national terms in the context of discussion of crime has nothing to do with changes in actual rates of crime by particular groups, and has everything to do with what people following certain political agendas (whether willfully or not) find it important to talk about. There has been no sudden change in the national origins of rapists during the time we have been writing this book, but there has been a dramatic change in the salience of a particular national origin in talk about rapists.
One thing that follows from the above data on word frequencies is that the resonances of crime talk include the salience of racial, ethnic, and national categories. These resonances are problematic. It is plausible that people who are regularly exposed to terms like “Black crime” have an increased tendency to form the idea that this is an important category in and of itself, and thus that there is an inherent link between being Black and having criminal tendencies. It is inevitable that for some this link will be essentialized, so that there is not only a false overattribution of crime to certain groups, but there is a view that this is an inherent property of people in those groups. Thus, it is plausible, and in line with his writing, that the resonance between crime talk and race led DiIulio to assume that the high rate of crime among Black youth was fixed, rather than fluctuating with social circumstances.
Mill’s reliance on the presence of a “real morality of public discussion” represents a serious limitation in his free-speech arguments, since it seems to demand of public speech standards that are not currently met, and for which we are unable to detect any clear positive trend in over 160 years since On Liberty was first published.
11.2. On Speaker Autonomy Let’s now consider a radically different argument for free speech than Mill’s, to see if it provides a justification for free speech. According to the speaker-based autonomy argument for free speech, a free-speech ideal is the direct result of each speaker’s individual liberty in a just society.
30 On the face of it, speaker-based autonomy arguments for free speech seem independent of the topic of harmful speech. But they are not.
In A Theory of Justice, John Rawls addresses the question of how a just state should go about ensuring liberties. He argues for two principles of justice. The first principle of justice is “each person is to have an equal right to the most extensive basic liberty compatible with a similar liberty to others.”31 If each person has a basic right to free speech, as the speakerbased autonomy argument would have it, it is at least limited to cases in which that speech does not restrict similar liberties from others. But some kinds of speech do have the capacity to reduce, impair, or downgrade the capacity of other citizens to act autonomously.
The speaker-based autonomy argument makes the idealization that all speech is straight talk. Hustle can manipulate the hearer, bypassing their rationality. We have, throughout this book, used the much-studied example of the way the word “welfare” in American English resonates with antiBlack racist ideology. When someone with anti-Black racist biases hears a program described as “welfare,” they tend to form a negative opinion of that program. It is the word “welfare” doing the work here, as the negative opinion tends to vanish when the program is described with other vocabulary. Typically, “welfare” is used as what Jennifer Saul calls a covert dog whistle; it operates by exploiting biases that are unconscious. As we saw in our introductory chapter, politicians exploit the racism of voters strategically by describing various programs they want to defund as “welfare.” Code words and dog whistles work in politics by manipulating citizens, often against their obvious material interests. White citizens regularly support defunding programs that materially support them, because politicians describe these programs as “welfare” or with other similarly coded vocabulary (e.g., “public”).
In Dying of Whiteness, Jonathan Metzl provides a book-length argument that the strategy of using vocabulary that resonates with anti-Black racism leads White voters to support policies that lead to mass death among these very populations of White voters.32 White voters support these policies because they are described with vocabulary that resonates with anti-Black racism, leading voters who have racial bias to adopt negative attitudes toward the programs.
Recall that covert dog whistles manipulate their audiences by operating on their unconscious biases. When one has been manipulated into a view in this manner, one’s adoption of that view is not autonomous. The assumption that all talk is straight talk idealizes away from speech that clearly restricts liberty. Politicians use covert dog whistles all the time to advance their agendas.
Even a perfectly nonracist but cynical politician could in principle use a covert anti-Black dog whistle to attract support from Whites with racial basis for his preferred policy position, by describing with dog whistles the position the politician wants them to oppose. Racist or not, politicians know this, and they use such manipulative tricks all the time (e.g., to get workingclass White voters to support cutting taxes for the wealthy).
The American journalist Christopher Rufo has led several successful propaganda campaigns. His first large-scale propaganda campaign was to recognize the dog-whistle power of invoking “critical race theory,” which he used to attack an array of targets, but especially public schools. In famous successive tweets, he wrote, We have successfully frozen their brand—“critical race theory”— into the public conversation and are steadily driving up negative perceptions. We will eventually turn it toxic, as we put all of the various cultural insanities under that brand category.
The goal is to have the public read something crazy in the newspaper and immediately think “critical race theory.” We have decodified the term and will recodify it to annex the entire range of cultural constructions that are unpopular with Americans.33 In this propaganda campaign, he directed the charge against federal bureaucracy generally, claiming that critical race theory was something like an official state ideology, and “is now being weaponized against the American people.”34 The goal here was to exploit unconscious racial bias as a weapon against public institutions (Rufo is a libertarian). Calling a program “welfare” will lead those with unconscious racial bias to think negatively of it. What Rufo discovered is that associating an institution with the label “critical race theory” taps into this same mechanism.
His campaign was massively successful. Educational gag orders have passed and are pending all across the country, forbidding K–12 schools, universities, and other institutions from teaching “divisive concepts,” any material that causes “discomfort, guilt, anguish or any other form of psychological distress.” Concepts such as “White supremacy” are banned by such bills, making it difficult to speak accurately about history. Multiple states have passed laws banning the teaching of Nikole Hannah-Jones’s 1619 Project, a multiauthor project documenting the centrality of slavery to all of the nation’s institutions. Rufo’s propaganda campaign, marshaling racial bias, used speech in the service of passing massive bans on speech.
In 2013, Russia passed its “Gay Propaganda Law,” which outlaws “promotion of nontraditional sexual relations to minors”; Russia’s attack on the LGBT community was part of its self-presentation as the world’s defender of “traditional values.” 35 The dog whistle used in country after country to target institutions in this now worldwide movement that Russia represents itself as leading is “gender ideology.” Hungarian Prime Minister Viktor Orbán used this charge to drive public sentiment against Budapest’s Central European University, which was eventually driven out of the country. Orbán also made it illegal for any of Hungary’s universities to teach “Gender Ideology.” More recently, Rufo has discovered the power of this label, and has launched a campaign against various institutions with its use. A New York Times interview profile of him writes, Mr. Rufo is convinced that a fight over L.G.B.T.Q. curriculums— which he calls “gender ideology”—has even more potential to spur a political backlash than the debate over how race and American history are taught.
“The reservoir of sentiment on the sexuality issue is deeper and more explosive than the sentiment on the race issues,” he said in an interview.
36 Speech can obviously be used strategically and manipulatively, exploiting unconscious biases. As Plato warned, it is particularly useful for those who have the character of tyrants—people whose only value is power.
Exploiting unconscious prejudices to one’s political advantage strengthens these prejudices.
Tyrants are ruthless in their pursuit of power. In the ruthless pursuit of power, strengthening preexisting prejudices is no barrier at all.
In addition to hustle, there is the problem of echo chambers. Recall the definition of an antagonistic ideological social group (section 6.6)—a community of practice whose discriminatory ideology is structured around a strongly negative collective emotional, dispositional, and attitudinal attunement to another group. A feature of antagonistic communities of practice is that they lead to divergent accommodation—reducing attunements to opposed groups.
Speech that fosters and maintains antagonistic social groups leads members of those groups to automatically dismiss the claims of opposing groups, to reject them just because of their group affiliation. Speech that fosters antagonistic communities of practice downgrades the capacity of other citizens to be heard.
Antagonistic echo amplification occurs when there are antagonistic communities of practice.
In such cases, there is little possibility for joint negotiation of meaning. People are essentially stuck in their echo chambers, as the mistrust that abounds prevents them from considering other attunements. Speech that leads to the formation of echo chambers creates obstacles to autonomy.
There are serious costs to departing from one’s antagonistic ideological social group, and in any case the trust required to acquire other attunements is absent in a highly polarized environment. Speech that fosters a necessary connection between one’s identity in an antagonistic ideological social group and a particular attunement—say, to a conspiracy theory—limits agency, as it places large nonrational pressure on members of that group to accept the conspiracy theory, or risk expulsion.
These facts do not mean that in a highly polarized environment, autonomous rational decisions are impossible. After all, people can and do reject views closely connected to their own social identities. But it’s important to bear in mind that the picture of autonomous rational agency is another idealization. It’s an idealization that underlies many if not all autonomy arguments for free speech. It is worthwhile to pause and assess its plausibility: Autonomous agent: “An agent who is sovereign in deciding what to believe and in weighing competing reasons for action.”37 We suspect this idealization underlies emotional attachment to particularly absolutist conceptions of free speech. Everyone has a strong urge to regard themselves as an autonomous agent. When everyone has a strong desire for something to be true, there will be a strong tendency toward wishful thinking. We have defined autonomous agency, as is standard, with the use of deciding what to believe and what to do, in short, in terms of accepting propositions. But in our framework, propositions are not special— attunement is the more general notion, and one can be attuned to propositions (belief), emotions, and practices. This naturally suggests a broader definition of autonomous agency: Broad autonomous agent: An agent who is sovereign in deciding what attunements to have.
The assumption of autonomous agency functions in autonomy arguments as part of an argument against speech regulation. Typically, one would argue that the proposed regulation of speech would limit agency. Agents, on this assumption, can pick and choose themselves what worldview to form from all the talking around them. So, limiting what options they are offered limits their agency.
However, the effects of speech are far broader than the offering of reasons. Speech can affect emotional state (e.g., producing fear) or social identity (e.g., strengthening some ties and weakening others). We have argued at length that many of the resonances of speech yield nondeliberative uptake, with our emotional reactions to speech being a primary case.
The claim here is not that we lack agency over our reactions to our emotions—we do have agency here. The claim is rather that we lack agency over the emotions we initially have toward speech we encounter.
The vocabulary of autonomy is not apt for expressive resonances. The case for nondeliberative uptake of emotional resonances is particularly strong.
It’s hard to see how a plausible claim of broad autonomy could be marshaled against the regulation of speech that harms through emotional impact and the intimidating expression of the power of a social group.
But even the narrow idealization of autonomous agency is suspect.
According to the idealization, humans can distinguish reasons that are compelling as such, independently from perspective, social identity, and other factors. Recall again the notion of an antagonistic ideological social group. Being in an antagonistic ideological social group places significant nonrational barriers on adopting opposing attunements, including opposing beliefs. In a polarized environment, we make many negative assumptions about people in opposing groups. Some of those assumptions will be falsely taken to be rational, rather than the result of our strongly negative emotional attunement to that group. The idea that people can just put aside their biases in a strongly polarized environment is fantasy. The idea that agents can consider opposing attunements solely on their own rational merits is inconsistent with the very real mechanisms that lead to echo chambers.
Even doxastic attunements that fit easily with our background ideology will be updated nondeliberatively.
Once we set the idealization of autonomous rational agency in a broader framework, with attunement, it is problematic—people often do not seem able to choose their emotional reactions to speech. It is only by narrowing the scope of agency to reasons that the idealization can even seem plausible.
But even thus restricted, the view that people always freely choose their views sounds more like wishful thinking.
Libel laws and defamation laws recognize other ways in which speech can restrict agency.
The scope of the free speech guaranteed by the speaker autonomy argument for free speech depends on an account of speech that provides (to reprise Rawls, above) “the most extensive basic liberty compatible with a similar liberty to others.” Insofar as restricting liberties is a matter of making it more difficult to attain them, this category of speech is much more capacious than the category of speech a free society would allow to be banned. It would, for example, clearly include code words and other kinds of hustle.
Similar concerns about subordinating speech practices also raise questions about autonomy-based arguments for free speech more generally, which tend to abstract away from the possibility of speech impairing or downgrading the power of autonomous action. In summarizing her critique of Ronald Dworkin’s autonomy-based argument for free speech, Susan Brison writes, What Dworkin fails to consider is that others’ rights, for example, their rights to free speech or to equality of opportunity, may be undermined by someone’s engaging in hate speech. Since he does not consider this case, he does not tell us how invoking the right to autonomy could resolve that conflict of rights. Furthermore, given that his theory of law specifies no procedure for weighing competing rights, this is not an oversight that could be rectified on his account.38 Let’s suppose that hate speech impairs the rights of others. If so, then autonomy-based arguments do not rule out restrictions on hate speech.
In his classic 1972 paper, “A Theory of Freedom of Expression,” T. M.
Scanlon provides a hearer-based autonomy argument for free speech.
Scanlon argues that restrictions on speech cannot be grounded in the potential harm speech has for leading to or encouraging false beliefs, or leading to or encouraging changes in preference. The basis for Scanlon’s argument is a claim about autonomy—“To regard himself as autonomous in the sense I have in mind a person must see himself sovereign in deciding what to believe and in weighing competing reasons for action.”39 Scanlon’s thought is that speech cannot be significantly curtailed without impinging on autonomy in this sense—the hearer must have the autonomy to decide what to believe and weigh competing reasons for action, and so speech must provide them with all the options.40 Scanlon’s picture of speech also presupposes the idealization of straight talk. Scanlon describes speech as straight talk, transparently offering the hearer propositions that she can freely accept or reject. But as we have seen, much speech is hustle. For example, simply asking leading questions can affect the actions of hearers, without them realizing it at all. Hustle is speech that can make a decision for its hearers. For example, in sections 4.2 and 4.3, we considered cases of psychologists, in effect, hustling experimental subjects through the presuppositions of frames; the subjects were coerced into making biased decisions or forming biased memories, but did not know they were being hustled. The first problem with Scanlon’s argument is that it presupposes the idealization that all speech is straight talk.
Scanlon’s argument also presupposes the idealization of an autonomous agent. As we have seen, a polarized environment limits or makes it altogether impossible to be one of Scanlon’s sovereign agents. Speech that fosters antagonistic ideological social groups and their attendant echo chambers curtails sovereign agency.
A central function of speech is to attune to social identity. Some social identities, in particular antagonistic ones, function to restrict or even render impossible developing attunements (including attunements to reasons) on their own merits, for example, on the basis of strength of evidence or of alignment with the agent’s interests. Speech that fosters antagonistic social identities destroys the healthy deliberative environment required for democracy, since negotiation between members of opposing groups becomes impossible. In a very clear sense, then, free speech may pose a threat to democracy.
The autonomy arguments against restricting speech fail because they idealize away from large categories of speech. They idealize away from speech that manipulates hearers (hustle). And they idealize away from the role of speech in creating and maintaining antagonistic social identities and echo chambers. They not only presuppose a healthy deliberative environment, but they also presuppose one in which all talk is straight talk.
In “Free Speech and Its Relation to Self-Government,” Alexander Meiklejohn defends free speech on the grounds that a democracy requires its citizens who decide issues to have “acquaintance with information or opinion or doubt or disbelief or criticism which is relevant to that issue.”41 For Meiklejohn, the function of speech in democracy is to inform. It is to educate the public about political questions. Jack Balkin summarizes Meiklejohn’s view as follows: In Meiklejohn’s model, free speech has constitutional value because it assures the flow of information that is relevant or potentially relevant to the democratic governance of a state. This formulation explains Meiklejohn’s famous comment that it is not important that everyone shall speak, but that everything worth saying be said. The receipt of information to the audience, and not speaker autonomy, is constitutionally valuable, because it allows the people to govern themselves, either directly through public debate and decision making, or indirectly through electing representatives and holding them accountable in elections.42 Meiklejohn’s free-speech defense has been critiqued for the limited scope of the principle—his defense does not extend to apolitical artistic expression, for example.43 But the problem is worse. In focusing on the informational function of speech, Meikeljohn’s free-speech defense would not protect speech that forms social identities, or speech that attunes to emotion. Yet these are central functions of speech. With our broader perspective, we can see that it is not just artistic expression that it fails to protect. It is social-identity formation. In “Cultural Democracy and the First Amendment,” Jack Balkin bases freedom on cultural freedom—the freedom to create meaning in your society with cultural and symbolic forms of production.44 This is much more promising by our lights. Unlike what Balkin calls “politico-centric” justifications of free speech that focus on its capacity for agential representation in the formation of laws, Balkin’s justification of free speech does not guarantee classical democratic equality—equal participation in the formation of laws that govern you.45 But it appears to guarantee full cultural participation and expression, and accentuates, rather than conceals, the cultural and symbolic force of speech. It clearly is focused on allowing the flourishing of social-identity expression.
A justification for freedom of speech based on cultural democracy faces the same pitfalls as its more “politico-centric” cousins. Social groups share not just views, but also practices and cultural expressions. Speech that creates antagonistic social groups will foster divergent accommodation to alternative cultural expression. Balkin’s argument for free speech abstracts away from expression that creates antagonistic social groups and echo chambers.
It too presupposes a healthy deliberative environment.
Other defenses of broad free-speech principles have more promise with dealing with the full reality of harmful speech. Judith Butler’s defense against hate-speech laws is premised on the view that hate speech harms, and that its harms are particularly difficult to overcome.46 Butler has several arguments against laws that ban hate speech. First, Butler is distrustful of a legal solution, pointing out that “the legal discourse in which the status of the performativity of hate speech takes place is its own performative exercise. In the current US political climate, the law that decides the question of hate speech tends to be applied inconsistently in order to further reactionary political aims.”47 We agree with Butler’s critique of law as a neutral arbiter. Indeed, we have argued for the incoherence of the notion of a neutral arbiter.
Butler’s most innovative argument against hate-speech laws appeals to her practice-focused approach to speech. Speech practices evolve and change. Butler argues that laws banning hateful expressions prevent positive linguistic change, freezing the harmful effects in place: “Keeping such terms unsaid and unsayable can also work to lock them in place, preserving their power to injure, and arresting the possibility of a reworking that might shift their context and purpose.”48 Butler argues that the best way to combat hate speech is to reappropriate it, as, for example, “Queer” was reappropriated by radical members of that community (so much so that it now figures as the name of a discipline, “Queer Studies”), thereby replacing a harmful practice by a practice that uses the slur to bond, and as an act of protest.
Butler’s intervention in the free-speech debate has been criticized on the grounds that reappropriation faces barriers that Butler failed to recognize.
The combination of a slur’s history, as well as the vicissitudes of its target’s identity and social location, can make it more or less difficult to reappropriate slurs directed against them.
Reappropriation is one mechanism to bring to bear to change the values and affect associated with a speech practice. As a specific mechanism for a targeted group to bond, it’s clearly too thin of a plank upon which to rest a full argument against hatespeech laws.
Butler’s defense of free speech is then, at the very least, incomplete.
We can see, however, that Butler’s defense against speech restrictions appeals to a tradition of rhetorical contestation, one that is embodied by, for example, the battle over the word “freedom” in US politics with which we began this investigation in chapter 1. On the one hand, we have a sense of “freedom” that resonates, as Toni Morrison urged us to recognize, with goods that are possessed in contrast to the state of Black Americans.49 On the other hand, we have a tradition of Black democratic political thought, from Frederick Douglass’s 1852 speech, “What to the Slave Is the Fourth of July?” to the present day, that emphasizes the rhetorical power of the democratic ideals, seeking to cleanse them of these racist associations.
Butler’s defense of free speech is essentially based on letting these rhetorical power battles over speech practices governing words play out.
What if these power disputes tend to resolve against the interests of oppressed groups?
Butler’s approach presupposes that groups will have the power to change practices. But who knows if this is plausible? Butler’s argument assumes much about the possibility of mutual negotiation of meaning.
The failures of arguments against speech regulation do not yield a positive case for it. Butler is right to emphasize that the state cannot be trusted to be a neutral arbiter. In the United States today, Rufo’s campaign against critical race theory has focused on banning speech that causes psychological discomfort to Whites, the dominant group. These laws ban concepts like “White supremacy,” making the teaching of US history either vexed or impossible. We can see the consequences of such bans in Russia, where laws forbid the teaching of the Soviet Union’s mass atrocities in the 1930s and ’40s, including Holodomor, the genocide Stalin caused in Ukraine.
It is a material reality, an on the ground fact, that many speech bans erase historical knowledge necessary for society to be self-governing. We see the consequences of such bans vividly, during the writing of this book, in the Russian public’s lack of response to Russia’s brutal and genocidal invasion of Ukraine. Many of the speech bans we see enacted today, what Timothy Snyder calls “memory laws,” target historical knowledge necessary for democracy. We should be against these kinds of bans because they ban knowledge necessary for self-understanding and self-governance.
We should promote an information space where mutual negotiation of meaning is possible.
This requires knowledge of the history of group relations, including histories of oppression.
There can hardly be the trust for mutual negotiation of meaning if there is one-sided ignorance of one group’s past history of brutal treatment of another.
A society based on the ideal of liberty is one that allows people to consider a rich variety of attunements, to freely form social identities. Achieving a society based on liberty is incompatible with the demonization of nonhateful alternative social identities. Speech regulation that is aimed at limiting the harm of antidemocratic social identities, or social identities that limit the freedom of others, is tolerable by the lights of our arguments.
Before offering a closing thought on the place of freedom of speech in democracy, let us briefly consider one final argument against speech regulation, already mentioned in our discussion of Butler, the question of “who decides.” Although we will not consider it in the same detail, it might be said that this argument is the most challenging of the three for anyone who, like us, would deny that there could possibly be such a thing as a neutral arbiter, someone who can act as a representative of the government and justly apply constraints on speech that are fair and equitable to all parties, without ideological bias toward one group or another. The argument has a pedigree. In his famous address to the English parliament on the value of free speech, the poet John Milton recounted a sorry history of censorship from ancient Greece to the Inquisition and beyond, pointing out its arbitrariness—“so often bad, as good Books were silenc’t.”50 Presumably his intention was to cast doubt on the reasonableness of imbuing any particular institution with the right to pass judgment on the value of words, even though he dared not directly question the ability of the august body he was addressing to make sage decisions.
The thought here is that there is no neutral arbiter for social identities. A troubling concomitant is that if we regulate the power of social identities that foment echo chambers and problematic antagonistic ideological social groups (say, genocidal ones), then those who have those social identities, when they take power, will regulate democratically healthy social identities.
So no speech regulation of any kind should be permitted.
The problem with this sort of argument is that it assumes those who advance illiberal social agendas will respect liberal norms if they have not themselves been highly regulated. History shows that this is folly. Those advancing harshly antidemocratic programs will regulate speech when they can, no matter what free-speech precedents liberals have established.
The rejection of liberal precedent and norms is in fact at the basis of these illiberal identities.
It might then be said that the question is not one of who will be a neutral arbiter, but of how we can avoid the very worst kind of arbitration.
11.3. Let Equality Ring Equality, in the words of Andrea Dworkin, was tacked on to the Constitution with spit and a prayer. And, let me also say, late.
—CATHARINE MACKINNON51 Given the failure of so many justifications for free-speech principles, what should we conclude? The paradox of democracy is just that—a paradox. It has been discussed for almost as long as democracy has existed. We will not solve it here. Our aim here has been to sharpen it, not solve it.
The two central values of democracy are liberty and equality.
Justifications of free speech have tended to draw support from the first of these values—liberty. But what if we drew, instead, on the second? Equality in democratic political philosophy does not mean material equality. It means, rather, political equality. It is standard to take political equality as the capacity to speak truth to power. A democracy should ensure that even its most powerless citizens have the capacity to speak truth to power. The capacity for powerful people to be embarrassed by being caught lying, for example, is a mark of a healthy democratic information space. The goal of free speech in a democracy should be to preserve a healthy democratic discursive space. This implies preservation of the capacity of journalists to operate freely and to have respect given to their findings. It implies preserving social identities that involve accountability and responsiveness to those who lack power, and, while not banning antidemocratic social identities, also not rewarding them. An equality-based defense of free speech would aim at preserving a democratic discursive space. But an equality-based defense of free speech should also accept that unregulated speech can lead to inequality; it can lead to speech harms that systematically wound and partially silence entire segments of society.
Democracy is a system for the development among the people of a nation of collective harmony concerning the government of those people.
Healthy democracies maximize the extent to which there is equal participation in this process, so that as much of the population as possible forms a single community of governmental practice. Political equality, then, is something that varies by degree: there is equality to the extent that a population has uniform access and power in the community of governmental practice, so that collective harmonization of those in the group is not at the expense of dissonance for those left out. Though any ideal of equality must deal with practical questions like how those without a voice (whether children, animals, or the planet itself) are represented, the pragmatic goal here is to support processes that lead to strengthening of the democratic discursive space, a public conversation in which more individuals become more equal participants in the democratic process of collective harmonization, and none become “more equal” than others. Our ambition should not be restricted to free speech, but should reach toward speech that is both free and equal.
As Lynne Tirrell writes, in an article focusing on the interrelated roles of speech regulation and counterspeech in combating misogyny: For anyone concerned with justice, the question is not whether something should be done about the misogynist onslaught girls and women encounter; the question is: What should be done, and who should do it? Supreme Court doctrine may favor counterspeech to tort remedies or criminalization, but to justify this we need a robust conception of what sorts of speech might have the power to counter oppressive speech, who can achieve it, and under what circumstances. In setting policy, we cannot assume a speech encounter between equally powerful adults, each fully free to speak their minds and each with the backing of deep and broad social norms. Where inequality reigns, the odds are not in favor of someone who tries to combat … bad speech of the powerful with … more speech of the vulnerable.52 This passage picks out one of many inequalities in contemporary society, inequalities that marginalize significant segments of the population so that their voices are suppressed while other voices are amplified. Democratic action is needed to repair an unhealthy democracy. Perhaps the major democratic force that has led to such repairs in the past century has been mass protest in the context of large civil rights movements.
An unhealthy democracy is one in which discursive access to governmental decision making is restricted to a community of practice that systematically omits segments of the population.
The higher the barriers to entry into this minority community of practice, the less democratic is the government, and if the minority community of practice that influences government decision making is restricted to a small elite, then there is no democracy at all.
When people are systematically omitted from access to governmental decision making, it is inevitable that the interests of members of the elite community of governmental practice will be better served than the interests of others. In such a situation, the system can only be maintained through coercion. The cheapest way to coerce is verbally, manipulating people through the use of intended hustle, especially undermining propaganda in which messaging presented as in service of an ideal in fact weakens practices that might favor that ideal.
Specifically, the elite may simultaneously present itself as the savior of democracy while actively working to destroy it, by convincing the people that inclusive democracy cannot work. Since democracy depends on large-scale collective harmonization, convincing people that democracy cannot work is most easily attained by (i) maximization of antagonism between subgroups, especially by convincing one group that it is existentially imperiled by another, and (ii) efforts to create big differences in discursive practices between groups.
As discussed in chapter 6, a combination of intergroup antagonism and intergroup differences in practice can lead to a runaway process of divergent accommodation, so that groups become less and less compatible with each other, and ultimately echo chambers form. An echo chamber, as we have seen, is a subgroup that has collectively harmonized around discursive practices and ideas that are incompatible with the discursive practices and ideas of other subgroups. Such incompatibility makes collective harmonization of the full population, and hence the possibility of a fully functional democracy centered on a common discourse, impossible.
Here we arrive at a conundrum: on the one hand, protest movements play a crucial role in supporting the push for political equality, but on the other hand, a protest movement, which has its own discursive practices, is always in danger of being isolated from the rest of the polity, becoming part of a problem producing antidemocratic fragmentation rather than part of a solution producing prodemocratic integration. Free speech is there to provide protection for such protest movements, so their voices can break through.
Right now, the most powerful forces in society, with the largest capacity to amplify their message, are focused on defending their power to guide public opinion by appeal to the ideal of free speech. Free speech, instead of being the mechanism by which protest movements can be heard, instead seems to be a tool to amplify the voices of the powerful, to protect them against critique. During the writing of this book, in the United States, politicians who present themselves as advocates of free speech are simultaneously passing bans on the teaching of concepts that are at the heart of protest movements for structural change, for example those associated with critical race theory. Banning such concepts ensures that the cases made by such protest movements will not be heard, that they will be consigned to an echo chamber, impossible to be heard outside its walls. An equality-based defense of free speech would not ignore threats to the speech of the powerful. But it would not center them. It would instead refocus attention on what it takes to give everyone a voice—a discursive environment that makes it obviously impermissible to ban concepts that enable the understanding of protests for structural change on behalf of disenfranchised groups. Only then it is possible for both mass protest movements and individuals operating within the electoral system and legislative process to speak truth to power.
1. MacKinnon, Feminism Unmodified, 207.
2. Meiklejohn, Political Freedom: The Constitutional Powers of the People, 26.
3. Plato, The Republic, 257 (565e).
4. Plato, The Republic, 259 (566e).
5. Mill, On Liberty, 9.
6. Mill, On Liberty, 51.
7. “Pornography’s Authority?” is chapter 4 of Langton, Sexual Solipsism. The 1993 paper is Langton, “Speech Acts and Unspeakable Acts.” 8. Delgado gives, for example, the case of Alcorn v. Anbro Engineering, Inc., in which a White supervisor accosted his Black employee who had confronted him with violating union rules with the N-word (Delgado, “A Tort Action for Racial Insults,” 97–98).
9. Lawrence, “Regulating Racist Speech on Campus,” 73.
10. Matsuda et al., Words That Wound, 8.
11. Maitra, “Subordinating Speech.” 12. To preserve a central role for authority in hate speech, some philosophers have compellingly urged that we need to broaden our conception of authority beyond traditional notions of epistemic or practical authority. For example, Michael Barnes has argued that we need the notion of a stable informal authority, which would explain why some propagandists’ assertions hold sway over large audiences, despite lack of markers of epistemic or practical authority, that is, no credentials or political office (Barnes, “Presupposition and Propaganda”).
13. Mill, On Liberty, 52.
14. Post, Democracy, Expertise, and Academic Freedom.
15. Post, Democracy, Expertise, and Academic Freedom, 34.
16. The argument for democratic competence is developed in Post, Democracy, Expertise, and Academic Freedom, chapter 2, “Democratic Competence and the First Amendment,” 27–60.
17. Hoffman, Race Traits and Tendencies of the American Negro, 99.
18. Hoffman, Race Traits and Tendencies of the American Negro, chapter 2, “Vital Statistics,” 33– 148.
19. Hoffman, Race Traits and Tendencies of the American Negro, 236–37.
20. Muhammad, The Condemnation of Blackness. He writes, “Using new data from the 1870, 1880, and 1890 U.S. census reports, the earliest demographic studies to measure the full scale of black life in freedom, these post-emancipation writers helped to create the racial knowledge necessary to shape the future of race relations. Racial knowledge that had been dominated by anecdotal, hereditarian, and pseudo-biological theories of race would gradually be transformed by new social scientific theories of race and society and new tools of analysis, namely racial statistics and social surveys. Out of the new methods and data sources, black criminality would emerge, alongside disease and intelligence, as a fundamental measure of black inferiority” (20).
21. Munton, “Prejudice as the Misattribution of Salience.” 22. Gould, The Mismeasure of Man, 65.
23. Fricker, Epistemic Injustice.
24. Collins, Black Feminist Thought, 253–57.
25. Dotson, “A Cautionary Tale.” 26. Hinton, America on Fire, chapter 2, “The Projects,” 46–69. She quotes, for example, Black Unity Conference spokesman Edward Davis: “if the chief can state there have been no problems in the black community, it is obvious that he is unaware of what is going on in his department” (58).
27. The following passage from Muhammad is striking in highlighting a specific change in the way crime was reported so as to focus on Black crime rates: “The nation’s most respected and authoritative crime source … simplified the racial crime calculus in 1930s America.
Blackness now stood as the singular mark of a criminal. ‘Negro’ became the only statistically significant category in the [Uniform Crime Report] tables upon which to measure ‘white’ criminality, deviance, and pathology” (The Condemnation of Blackness, 270). 28. DiIulio, “My Black Crime Problem, and Ours.” 29. All frequencies cited were obtained using Google Ngram searches on October 18, 2022, and relate to Google’s American English books corpus, with years 1969–2019.
30. In “Freedom of Expression,” Joshua Cohen argues against this view, on the grounds that it roots the ideal of free expression in a more controversial ethical doctrine, that we have fundamental liberties.
31. Rawls, A Theory of Justice, 53.
32. Metzl, Dying of Whiteness.
33. The tweets were for a while unavailable through Twitter. As of October 24, 2022, the tweets are available at https://twitter.com/realchrisrufo/status/1371541044592996352.
34. Rufo is cited in Laura Meckler and Josh Dawsey, “Republicans, Spurred by an Unlikely Figure, See Political Promise in Targeting Critical Race Theory,” Washington Post, June 21, 2021.
35. Tat Bellamy-Walker, “Russian Court Dissolves Country’s Main LGBTQ Rights Organization,” NBC News, April 26, 2022, https://www.nbcnews.com/nbc-out/out-news/russiancourt-dissolves-countrys-main-lgbtq-right s-organization-rcna25874.
36. Trip Gabriel, “He Fuels the Right’s Cultural Fires (and Spreads Them to Florida),” New York Times, April 24, 2022, https://www.nytimes.com/2022/04/24/us/politics/christopher-rufo-crt-lgbtqflorida.html.
37. Scanlon, “A Theory of Freedom of Expression,” 15.
38. Brison, “The Autonomy Defense of Free Speech,” 325.
39. Scanlon, “A Theory of Freedom of Expression,” 15.
40. John Milton similarly argues that the dangers of free speech, and particularly the possibility of pernicious ideas, do not justify censorship, because encountering pernicious ideas is in fact necessary for developing both our rational faculties and a virtuous character, as well as properly grounded conviction in the truth: “He that can apprehend and consider vice with all her baits and seeming pleasures, and yet abstain, and yet distinguish, and yet prefer that which is truly better, he is the true warfaring Christian. I cannot praise a fugitive and cloistered virtue, unexercised and unbreathed, that never sallies out and sees her adversary, but slinks out of the race where that immortal garland is to be run for, not without dust and heat” (Areopagitica, 18).
41. Meiklejohn, Free Speech and Its Relation to Self-Government, 26.
42. Balkin, “Cultural Democracy and the First Amendment,” 164. 43. See, e.g., Chafee, review of Free Speech and Its Relation to Self-Government; Kalven, “The Metaphysics of the Law of Obscenity.” Meiklejohn famously argued in response that one could incorporate art into his conception of free speech, because poetry, movies, and other cultural production ultimately aid political decision-making (“The First Amendment Is an Absolute,” 245).
This is, however, an unfounded assumption. There is art that can undermine reason in characteristically propagandistic ways.
44. Balkin, “Cultural Democracy and the First Amendment,” 1054.
45. Balkin, “Cultural Democracy and the First Amendment,” 1056.
46. Butler, Excitable Speech. The book begins “When we claim to have been injured by language, what kind of claim do we make? We ascribe an agency to language, a power to injure, and position ourselves as the objects of its injurious trajectory.” (1) Later, she writes, for example: “If we accept that hate speech is illocutionary, we accept as well that words perform injury immediately and automatically, that the social map of power makes it so” (101–2).
47. Butler, Excitable Speech, 39.
48. Butler, Excitable Speech, 38.
49. Morrison, Playing in the Dark, 64.
50. Milton, Areopagitica, 28–29.
51. MacKinnon, Feminism Unmodified, 207.
52. Tirrell, “Toxic Misogyny and the Limits of Counterspeech,” 2438. Here Tirrell’s argument for centering issues of equality in determining how competition between speech interests should be resolved echoes a theme in feminist legal theory that is clear in MacKinnon’s work: Any system of freedom of expression that does not address a problem where the free speech of men silences the free speech of women, a real conflict between speech interests as well as between people, is not serious about securing freedom of expression in this country. (Feminism Unmodified, 193)