Beyond “Mention vs. Use”: The Linguistics of Slurs

There is a group of people who are extremely concerned that they might get in trouble for saying the N-word.

I recently had the pleasure of speaking with Eiynah, the host of the podcast “Polite Conversations,” about this concern. Since we’ve both been keeping our eye on the Intellectual Dark Web, we’ve noticed that a particular favorite topic of certain members is finding stories where someone got in trouble for using the N-word in the workplace and using that as an example of the ways woke mobs are destroying academic freedom. We had a conversation about this on her podcast (listen to it here), and I created this essay as a way to link to some of my sources and flesh out the conclusions we arrived at.

The N-Word in the Culture Wars

If the free speech warriors are to be believed, the world has lost track of the mention vs. use distinction: the idea that using a slur as an epithet is completely different from quoting it or referring to its use by others. Two examples that we’ve seen being pushed over and over again are the creative writing professor who was investigated for quoting James Baldwin, and the business school professor who said a word in Mandarin that someone thought sounded like the N-word. In both cases, a student complained, the school investigated, and the professors were cleared.

But I thought the puritannical refusal to allow the word’s mention was sweeping the nation and ruining lives! Peter Boghossian of the Sokal Squared hoax says it’s causing institutional collapse, for goodness’ sake. But if that were the case, why was Laurie Scheck cleared? She was investigated and found to have used the word in a justified way, in proper context. So apparently the horrible thing that is going to bring about an academic apocalypse is… investigations that come up clean. Do these free speech culture warriors want a world where students never raise objections that ultimately result in the investigated party being cleared? Seems implausible. There are about 20 million people in college in the U.S. and they all have different values and experiences. Some of them are going to make complaints some of the time. Do they want a world where universities dismiss complaints without checking into them? That would be unethical, to say the least.

But to these figures, the mere idea that there are times when referring to the N-word (by actually saying it, not by saying “the N-word”), even without “using” it, might also be a troubling choice in some cases, is absolutely beyond the pale.

So linguists in the public eye seem to be selling the idea that objecting to the uttering of a slur at work is a failure to understand that saying it in a “referring” way is perfectly okay. Feels a little simplistic for a linguistic analysis, but maybe that’s just me. Seems like it would be helpful to define some terms. What are the linguistic features of a slur? Are all slurs equally impactful? Are there any times when mentioning a slur is also worthy of professional or social consequences? These are not questions addressed by the anti-woke linguistics celebrities, even though it is ostensibly within their field of expertise to discuss attitudes about language from a scientific perspective. Instead, we get hot takes about the Woke Elect or Orwellianism or struggle sessions or whatever, some advanced by the linguists and some by others, but amplified and endorsed by these linguists. We get jabs like this one, where McWhorter decides to mock the idea that lefties tend to frame bigotry as an issue of impact rather than intent, apparently forgetting that there’s a well-established concept in linguistics called “perlocutionary force” that kind of suggests that impact can in some ways be separated from intent.

Set aside the fact that academic freedom in universities is not threatened by the left, set aside the fact that the martyrs these figures choose are never actually punished (or if they are, they are guilty of way more than mentioning a slur), and just marvel at the sheer volume of opining about how it’s actually just fine to mention the N-word (by saying the full word) as long as you’re not actually calling anyone an N-word.

In a world in which human beings are mortal and have a limited amount of time to consider and solve problems, this is where their time is going. What does that say about their priorities? Rather than asking the interesting questions associated with how to treat slurs — and taking the opportunity as high-profile scholars to direct people to the work of other linguists — they take an opinion-driven, political approach, working overtime to convince readers that free speech on campus (and elsewhere) is under attack by “the woke” and that random college students taking issue with their professors’ word choice is basically the same thing as what Orwell was warning us about in 1984. Remember when Winston was fired from his job for mentioning a racial slur to his employee? Neither do I, but it must be in there somewhere.

So what are the more interesting linguistics questions that these figures could be exploring instead of trying to force it into the larger “cancel culture” political argument? I’m glad you asked.

Can Words Do Real Things?

On August 8, 2016, Steven Pinker tweeted, “the first insight of linguistics, going back to Plato, is that words are conventions, without magical powers. That’s being nullified by PC/SJW attacks on mentioning taboo words, even ironically or in works of art.”

In this tweet, he linked to an essay written for Quillette (yikes) by right-wing essayist Heather Mac Donald (yikes yikes) about why nobody should question someone using the N-word in a “mention” context as opposed to being hurled as an epithet. In it, she misrepresents the work of linguist Taylor Jones, who published a reply on his blog. The primary argument in Mac Donald’s essay was that people just mentioning the word have their lives ruined over it when they did nothing wrong. We will get into why that’s silly on a linguistic basis, but as for the claim that “[i]ts mere presence in the mouth of a white person launches a nuclear bomb against blacks; the transgressor will be punished accordingly,” well, the continued employment of Laurie Scheck is evidence that justified mention of the word will not ruin your life.

But let’s discuss Pinker’s claim in his tweet promoting this pile of right-wing resentment dressed up as political analysis. Pinker, who is not a linguist but plays one on Twitter, enraged a significant number of Too Online Linguists such as myself with his little intro when he said “words are conventions, without magical powers.

*J.L. Austin Has Entered The Chat*

The idea that words have a material impact goes back to the beginning of words, but in white western linguistic tradition, we have the advent of the field of pragmatics, first introduced by Ferdinand de Saussure in the late 1800s. Pragmatics endeavors to explain how meaning can be created using not only what the conversation participants know about the literal meaning of the words or signs in an utterance, but what they know about the participants, the time and place of the utterance, the inferred intention, and the history of the words and phrases at play, among other things. On top of this, certain utterances can have performative elements, as described in J.L. Austin’s work (e.g., “How To Do Things With Words”). A performative isn’t like performing in a play, it’s more like performing surgery or performing a backflip. It means making something very real happen just through your own actions. Consider a judge in a courtroom. The plaintiff and defense have given their case. The jury has returned a guilty verdict. The judge says, “plaintiff is hereby sentenced to serve eighteen months in prison.” Those words, spoken by that person, in that context, are not just words. They set a series of events into motion, events that will severely impact the lives of multiple people in that room.

Now, imagine a proposal. One person goes down on one knee, maybe at a restaurant where spectators are around. The other person in the couple freezes, stammers, then says, “Oh no, no I can’t do this.” What is happening in the mind and body of the person who just got rejected? Dunno, probably nothing. Hey, it’s just words, right?

Despite the protestations of our rationalist friends, emotion is a fundamental component of cognition. When I asked other linguists about the material impacts of language, sociolinguist Kelly Wright directed me to a chapter in the Oxford Handbook of Psycholinguistics, by Jos van Berkum. In it, he says:

The chapter includes a lot of really fun diagrams, too, such as this one illustrating how an emotion causes a “processing cascade” in someone who appraises some stimulus as relevant to their interests or concerns:

So the emotions a person feels upon hearing certain words will affect how they process language and will impact their motivations, physiology, cognition, and behavior. Not only that, but the emotional content will impact how they even process the language they’re receiving. What was that about words not being magic, Pinker?

So, the IDWers are choosing to completely ignore the performativity and emotional impact of language, but their mention vs. use obsession also obscures the specifics of how slurs work, and beyond that, how the N-word works. As everyone who works on slurs knows, the N-word is the most extreme example. So what does that mean for its use?

What Makes A Slur?

There are two primary schools of thought regarding what makes a slur, a slur. One is that a slur is any term for a group of people that becomes associated with bigoted abuse. As Nunberg (2018) says, “racists don’t use slurs because they’re derogatory; slurs are derogatory because they’re the words that racists use.”

Another way to define what is or isn't a slur is to take the semantic view. In semantic explanations, a slur is defined as a coherent class that satisfies some criteria. According to Davis and McCready, the criteria are:

As someone who studies the properties described in (iii) I would argue that those properties are not intrinsic in the sense that they are stable, objective categories, but they are properties that our society considers to be intrinsic. After a conversation with one of the authors of this paper, I believe they agree. We do not consider things like political affiliations or opinions to be intrinsic properties: by this model, “nazi” is not a slur, nor is “TERF.”

In the pursuit of a semantic explanation, philosophers of language have largely coalesced around a conception of slurs where you take a non-pejorative word for a group of people and then you “just add bad,” as philosopher Jennifer Foster helpfully describes it, meaning you insert the implication that it’s bad to be a member of that group. That means derogatory classifiers (some of which are slurs and some of which aren’t) have non-pejorative words that mean roughly the same thing without the hateful or contemptuous connotation. In her paper, “Busting the Ghost of Neutral Counterparts”, Foster compiles some examples:

One problem is, I think most people (all people? hopefully all) would agree that there’s a huge difference between calling someone a ‘treehugger’ and calling someone the N-word. In short, one is an insult and the other is a slur. It sucks to be insulted but it’s not the same thing as being called a slur, for reasons you may or may not be able to articulate fully. The discussion of what separates them is where a lot of the interesting meat is, and is one of the ways our linguistics-adjacent celebrities are completely failing us. What separates insults from slurs isn’t necessarily answerable by the semantic argument, but the pragmatic argument that slurs are slurs when they become associated with bigoted abuse might get closer to it. We’ll look at the ways that happens in the next two sections. But first, let’s finish Foster’s critique of the “just add bad” view.

In accordance with the “just add bad” approach, the pairs of words aren’t perfect synonyms, since the derogatory classifier always implies the person using that word is taking a negative stance toward the people it describes. A slur for group n means “member of group n but also I believe they are contemptible by virtue of belonging to group n.” But there’s more to it than that: Foster argues that the meaning of the slur and its non-derogatory equivalent don’t map perfectly, they only overlap, since she’s found examples of people saying you can be an NPA without being a DC, and vice versa.

If those sentences are allowable (and they must be, since Jen found them in the real world!) then it’s not just that a slur is when you invent a synonym and also add the bad to it. There must be more to people’s understanding of the slur than just the shared semantic truth-conditional meaning. Take a look at these examples:

The fact that (70) makes sense but (74) doesn’t suggests that by saying “Whatever, they’re stilll ch*nks,” A recognizes that the Vietnamese people they are objecting to don’t strictly meet the truth conditions of the NPA “Chinese person,” but that they don’t consider that a dealbreaker for calling them the slur. Weird! Foster argues that the willingness of bigots to make exceptions, to overgeneralize when convenient and exempt group members they consider “one of the good ones” when they can be useful, is evident in this practice of not treating NPAs and DCs as perfect synonyms. Traditional definitions of the various types of bigotry as a hatred of a particular population fail to recognize these strategic exceptions and inconsistencies and, as Kate Manne points out, how maintaining that they actually like certain members of that population helps insulate them from accusations of bigotry.

That’s why the researchers named above believe the strict semantic definition of a slur (as a term for a population with the added meaning that its user considers members of that population to be contemptible just for being a member) is inadequate, and we have to look at the use and impact of slurs. Fortunately, pragmatics is able to help us understand it.

Two Elements of Slurs: derogation and offensiveness

Most linguists looking at slurs agree that the meaning of a slur depends on two elements: derogation and offensiveness. A word is derogatory if the person using it endorses the content that it is contemptible to be a member of the group being referred to. A word is offensive if it invokes a sociohistorical meaning that involves extreme bigotry, horror, or violence.

Since derogation reflects the intentions of the speaker, it can be mitigated by the person who is saying the word. You can say a slur without attributing the feelings normally associated with it to yourself. Imagine you’re walking with a friend and someone shouts a racial slur S at them. You might turn to your friend and say, “I can’t believe they called you an S!” In that case, you aren’t the one calling them an S, you’re just pointing out how surprising it is that the other person said it. Similarly, you might need to quote someone using the word in order to convey how toxic their ideas were. Maybe you want to explain to someone how homophobic Nixon was, so you decide to remind them of this quote:

The context and your intent will show that you don’t endorse Nixon’s use of the two slurs in this quote. This means your uttering it won’t be considered derogatory.

The part you can’t avoid is the offensiveness, and that’s because of a process called invocation, which happens regardless of your attitude toward the slured group. It’s just like invoking the spirit of the mummy from the masterpiece 1999 classic, The Mummy.

When Evelyn read the words in the Book of the Dead, she didn’t know it was going to do anything, she just figured she was doing research. But the curse didn’t care about her intentions. All that mattered was that the words were said. That’s basically how invocation works with slurs. Invocation is independent of the goals of the people in the conversation.

Invocation is a term in pragmatics that means there are certain facts, histories, and ideologies that all parties are at least somewhat aware of, which will be associated with a slur no matter what the intentions of the person saying it. Remember the diagram showing how we process emotionally competent stimuli? The history of the word will come to you no matter what you think the other person meant. The more violent and vitriolic the sociohistorical content, the more justification you will need for saying it, and the more work you have to do before you say it to ensure that the people around you are prepared to hear or see it. Even so, you can’t spare them the effects of invocation.

So we know it’s possible to counteract the derogatory nature of the word, by making it clear that you don’t endorse its use. That’s what people mean when they say mention is different from use. But because a slur is always offensive, because it always invokes information that you can’t negate, then the recipient draws inferences about why you chose to utter the word. That means even if you did all the work required to distance yourself from the derogatory sense, if it wasn’t strictly necessary to pronounce the word itself, the recipient will make inferences about your motivations that will undo your efforts to mitigate the derogation.

This is where it becomes clear that there must be some kind of sliding scale. You probably sensed that the calculus would come out differently if you needed to say something like f*g or ch*nk, versus something like “trailer trash.” Even here, I blanked out letters in the first two but not the other. Why is that? Because the content it invokes is not as intense, the cost of saying it isn’t as high. Anderson and Lepore investigate this question in “Slurring Words.” When a word is so deeply connected with extreme violence, systemic inequality, and serious injustice, it becomes phonologically (or graphically) poisoned. This is why the N-word especially is an edge case, and why it really stretches the explanatory power of the “mention vs. use” distinction that our celebrity academic friends so love to cite.

So what happens when you choose to say it in a context where the audience doesn’t accept that it was important enough to justify saying the actual word?

Mentioning the Slur: Deciding To Be That Guy

The linguistic literature is firm that choosing to say the slur when there’s a non-derogatory option, a redacted version, or a euphemism available says something about what kind of person you are identifying as. Nunberg disagrees with Henderson, Klecha, and McCready about where that line is. Reasonable people can disagree on this, but it’s not reasonable to claim that there is no such thing as a “mention” case that’s out of line.

Another split in the linguistic literature is whether slurs are best described by thinking of them as mixed expressives or by using Gricean models.

An expressive is a word or phrase that serves to convey your attitude toward whatever you’re talking about. It’s also performative, because it has strong emotional content that will impact the recipient. Take, for example, a sentence uttered at a house party:

“Seth brought his fucking guitar.”

The addition of the expressive “fucking” here illustrates a whole host of stances, or attitudes, about the fact that Seth has brought his guitar. He’s going to grind the whole party to a halt so he can struggle through Wonderwall in the vain hope of charming some sexy person, to the detriment of everyone else at the party.

Because the emotion and historical use of those performatives are part of their meaning regardless of whether they’re being said in earnest, quoted, or even said by accident, the effect is the same and mention vs. use doesn’t apply.

This seems like too hard a line, though. If that were the case, the professors who were investigated would have been in more trouble. Seems like there are times when it’s necessary to reproduce the word, and those times

Some linguists argue that a slur is a mixed expressive: it carries at-issue content, meaning the semantic content of the slur (there is a group of people with some set of characteristics), and expressive content, meaning facts, attitudes, and prejudices that exist outside the speaker. This should sound familiar; it is essentially derogation vs. offensiveness.

The Gricean model goes back to H.P. Grice, who developed the cooperative principle of pragmatics. The idea is that people engaged in communication want to understand each other and will do what they can to ensure that understanding is reached. He proposed four maxims of communication, or assumptions we all have when interacting: quality (people will tell the truth), quantity (people will tell you as much as you need to know), relevance (people won’t go off topic) and manner (people will use appropriate words, speaking/signing style, speak/sign in order, etc.). When someone obviously violates one of these maxims, the recipient will draw a conversational implicature, which means they will make inferences about the person’s intentions when they spoke in an unexpected way.

Using a slur or any taboo language is a violation of the maxim of manner, so it tells the recipient that something else is going on: the speaker is comfortable with the offensive content of the word for some reason. If the justification is clear and satisfactory (like in the case of Scheck doing a pedagogically necessary discussion of the word’s use by James Baldwin), most recipients will consider it acceptable. If the justification is shaky, then it reflects poorly on the person who produced the slur. Once a word is understood to be a slur, using a word that’s so tainted by its history means you consider yourself the type of person who is comfortable invoking the content associated with that slur.

Most people can imagine that there are times when someone is absolutely on the right side of the mention vs. use distinction but there’s just no real justification for using it. Take a look at this tweet from popular free speech warrior Stephen Knight:

According to both the expressive explanation and the conversational implicature explanation, just tweeting, “N****r is a racial slur” just looks like someone wanted to say the N-word without getting in trouble. There’s no justification there, there’s no specific incident he’s criticizing or person he’s exposing as racist, he’s just typing it to type it. And the fact that everyone knows that n****r is a racial slur means that stating it clearly violates the maxim of quantity by giving unnecessary information. That means if you’re looking for conversational implicatures, Stephen doesn’t come off looking too awesome.

Now, imagine Stephen wrote or said that, apropos of nothing, to someone he employs, or to someone he’s in charge of teaching and evaluating. Now it’s even weirder and more worth looking into. That’s starting to look like someone who can’t be trusted to be professional and create a healthy workplace/academic culture.

Whichever explanation is right, or if there’s a third, better explanation (I think Social Positioning Theory is a good direction to look in, but then again that’s what I use so I’m biased), it very well may be that people looking to decide when there should be consequences ought to use three categories for understanding the production of slurs, not just two: use, unjustified mention, and necessary mention. This is why I say that the IDW figures who are obsessed with “academic freedom” (but are strangely silent when it’s not white conservatives or people who said a slur) are seriously simplifying the situation. Just like with cancel culture as a whole, there are situations where all of these guys think consequences are justified, they are just clinging to a simplistic and outrage-machine-fueling line to raise their own profile as The One Who Tells It Like It Is. And it is especially frustrating that some of them are linguists but they never want to talk about the research being done on this.

Tone Policing And Epistemic Injustice

When I asked my linguistics and philosophy friends what they thought was relevant to the question of racial slurs and how they function in American societies, Megan Figueroa of the Vocal Fries podcast brought up the asymmetry of civility and tone policing. In fact, this tendency is actually quite well-described by feminist philosophers. For those of us who researched the sequence of events after the open letter to the Linguistic Society of America went out, it could not be more painfully obvious that there is a different standard for what constitutes civility for folks on either side of the “cancel culture” debate. Pinker and his allies got to call us “out to lunch,” compare us to Maoists, Big Brother, demons at the gates, and all kinds of other hyperbolic epithets, but our meek suggestion — that Pinker’s tendency to contradict and obscure the work of less prominent linguists in pursuit of an anti-social-justice ideology might make him an unsuitable representative for us — was completely over the line. When Obama went on the WTF podcast and said the N-word, certain people hated it. It seems like when you’re calling out bad behavior in those who fight on the side of the status quo, the mere fact of your objection is considered uncivil. But when you’re the one standing athwart history, yelling “Stop,” when you’re gainfully employed regularly publishing op-eds and books about how social justice is the Real Threat to Western Civilization, then it feels like a natural right that you should get to say the N-word whenever you want.

I guess I should have predicted that I would be on the receiving end of this exact asymmetry of tone policing. By appearing on a podcast to talk about racial slurs, where the host exposes IDW figures who have said them for no good reason other than to rage against the idea that they shouldn’t, I’m a natural target for the same people who love to talk about mention vs. use being the only relevant principle. It’s literally what I described in the previous paragraph. Come on, Caitlin! Think!

Sounds like they’ve created a strawman argument for the “woke” boogeyman they imagine for linguists like me. We do not say that there’s never a situation where uttering or reproducing a recording of the word is justified. We only ay that not all mentions are acceptable, and the judgment for when they are is context-specific. Based on the arguments I’m seeing from the IDW (and Pinker and McWhorter), they think it’s never okay to question someone’s mention of the N-word.

Unfortunately for those figures, they live in the same world as we do, and there’s no magical shroud of invincibility they can wear to avoid the fact that others will perceive them as the kind of willfully insensitive person who is looking for excuses to say a word that they know is a slur.