The times we live in are starting to seem disturbingly biblical. Worldwide, there have been fires raging across continents, unbearable heatwaves, and unseasonably warm winters. There’s been major social upheaval, too, from strikes, protests, and elections. And now, the world suddenly finds itself in the midst of a possible pandemic that dare not speak its name.

In a very short span of time, cases of a new coronavirus originating in Wuhan, a city in the Chinese province of Hubei, have spread exponentially around the globe. The virus has caused sickness and death in such unmanageably large numbers that public health systems in some countries, lacking resources, have been strained to the breaking point. In the early confusion and growing public panic, as researchers raced to investigate the makeup of this strange new virus, amid infection and fatality rates much higher than the seasonal flu, there was scarcely time to worry about the seemingly tangential linguistic implications of when to call a pandemic a pandemic and what we should consider when naming a new disease.

As with many other recent global crises, there has been a flood of dangerous misinformation, and with it, the insidious “viral” spread of toxic speech through social media. With people beginning to grapple with fearsome unknowns over which they have little control, the public conversation has nourished a sense of permissiveness around acts of outright racism and physical violence. The wrong kind of name might give listeners the wrong idea.

The path between the so-called linguistic dog whistles and the violent action they can inspire is a murky one. How do words that should be innocuous, like “poor,” “welfare,” or “inner city,” or perhaps more topical words, like “socialist,” “illegals,” and “establishment,” go negative?

By February 11, when the World Health Organization gave the name COVID-19 to the mystery disease, it was perhaps too late. A range of colloquial labels had already taken root. Since there’s more than one kind of coronavirus, people came up with intuitive names that referenced the disease’s origins, such as the “China flu” or “Wuhan virus.” While these names may have been used neutrally at first, it became clear that they had the potential to reinforce harmful stereotypes. As a disease name, COVID-19 is easy to pronounce and also follows the WHO convention of not referencing animals, people, or places, to avoid stigmatizing them. (The WHO’s now-official name for the virus is severe acute respiratory syndrome coronavirus 2, or SARS-CoV-2; the disease that the virus causes is coronavirus disease, or COVID-19.)

Still, the association of COVID-19 with China and Asian people in general has lingered. Chinese-owned businesses far from any outbreak felt the pinch, with their customers staying away, and many Asian people are reporting a rise in racism. It’s clear that in the right context, there’s a hidden power associated with a name, despite what Shakespeare might say. Taiwan, in a tussle over nationhood with China, was controversially labeled by the WHO at times as “Taiwan, China,” or “Taipei” in its outbreak updates. Taiwan decided to continue using the earlier term “Wuhan virus” widely, ostensibly to avoid confusing the public with too many names. But perhaps there was another reason for not using a naming convention that most other countries quickly adopted.

How does a name that perhaps arose out of expediency develop negative connotations for a wider English-speaking audience?

Being associated with illness, much less a pandemic-causing virus, is not the most positive of things. As Lawrence Besserman notes, being ill in the English language is no fun. Metaphorically, illness is often conceptualized as a personal failing, as we blame the sick for actively catching illnesses rather than having illness happen to them. Although “ill” is not etymologically related to “evil”, Besserman shows how the two became synonymous, as “ill” attracted more negative metaphorical connotations of moral depravity, just as “sick” similarly took on the pejorative senses of being mentally ill or even “perverted.” These senses remain in terms like ill-advised, ill-bred, ill-tempered, ill-will.

But more than that, as the recent incidents of coronavirus-related racism show, many countries, in particular the U.S., have a long and unhealthy history of associating germs with the idea of being foreign. The long-held stereotype of unclean immigrants importing germs from strange places resulted in invasive public health practices before immigrants were allowed to enter the country, with many being deported for spurious health reasons. Disease was seen as related to criminality, poverty, addiction, immoral behavior, even communism. Those same germs, of course, don’t particularly care for borders; they could have just as easily hopped on an unsuspecting American coming back from a vacation abroad.

It’s unsurprising, then, that the name of a disease, if it incorporates the name of a people, will likely eventually result in stigmatization, especially given the gravity of the current outbreaks. So much so that Chinese officials are reportedly trying to push the false story that COVID-19 may have been found in China, but didn’t originate there.

In today’s heavily charged political environments, more and more people are aware that language can be weaponized and used to frame narratives that aim to persuade people one way or another. But when we think of weaponized language, often it’s a pretty blunt instrument. We tend to focus on the obvious slurs that a speaker and a wider audience can understand are clear-cut negative terms that convey bigoted intentions and aim to wound.

We live in a world where it’s much harder to use outright slurs as linguistic weapons in a public or professional sphere, because racism is widely seen as unacceptable, and calling someone a racist is still considered a serious accusation for which there’d better be evidence. When slurs are used, it’s easier for observers to call out racism with certainty, and harder for those who use slurs to plausibly deny it.

Perhaps a more harmful kind of speech lives in a gray area, merging and intertwining with norms and nudging them in certain directions, but having plausible deniability when confronted. Right-wing Dutch politician Geert Wilders once staged a scene where he asked a crowd “Do you want more or fewer Moroccans?” which resulted in anti-Moroccan chants from the audience. After being convicted of hate speech, he denied it, plausibly, saying, “I cannot believe it, but I have been convicted because I asked a question about Moroccans…. The Netherlands have become a sick country.”

It can be hard to tell when speech is intentionally coded to get listeners to infer something that’s never directly said—and when it’s really what it says on the surface. Last year, when Joe Biden, a presidential candidate known for his support of African-American communities, said “poor kids are just as bright and just as talented as white kids,” what did he mean? Perhaps it was a gaffe, or perhaps it was an unwitting code that drew on assumptions that some of his listeners shared. Unlike a slur, it’s hard to tell what his intentions were. Crucially, though, intent matters less than whether an audience can still understand and receive the message as coded, and in some cases, act on it.

When Barack Obama suggested during the 2008 presidential campaign that his opponent, John McCain, and the Republican Party would push the fact that he “doesn’t look like all those other presidents on the dollar bills,” McCain’s campaign responded furiously: “We are not going to let anybody paint John McCain, who has fought his entire life for equal rights for everyone, [sic] to be able to be painted as racist.” How did McCain’s campaign read fighting words into Obama’s statement? For those who are looking and are of a similar mindset, the answer is obvious, but for others, it might not be. This is why these political code words are often called dog whistles: those who are sensitive to it can hear it, but those who aren’t, can’t.

This gray language can spread across a community and infect the semantics of words with a new toxicity, if listeners can make the metaphorical connections and choose to endorse them by not challenging them. The more they’re repeated in those fuzzy-but-negative contexts, the more these coded meanings become normalized and accepted. But for philosopher Lynne Tirrell, words don’t go negative just through frequency of use in a negative context. Conversation is a joint, cooperative activity, like a dance, with actions and reactions. It needs listeners to react, by accepting or resisting what’s said. For toxic speech to evolve new inferences and meanings, rather than challenging or resisting the semantic change, an audience has to endorse it by taking it up and using it themselves, spreading it among the community like a virus that leaps from host to host. It’s in this way that media can unwittingly spread biased political talking points, by dutifully reporting toxic speech without commentary or challenging false information in a distorted sense of neutrality.

Tirrell considers the idea of an epidemiology of toxic language, conceptualizing language as something that can have the power to cause harm, by degrees, just like a poison or a disease. When we think of speech acts in terms of toxins or viruses, it’s a useful medical metaphor that helps us makes sense of how speech, even when it doesn’t contain any slurs, name calling, hate speech, or other derogatory language, can nevertheless be harmful, even to the point of violence, especially the further and more frequently it’s spread around.

“Words have killed my country,” says Dr. Naasson Munyandamutsa, a Rwandan psychiatrist who survived genocide. Tirrell describes how Munyandamutsa was referring to the derogatory words that were once used to refer to Tutsi guerrilla soldiers, words like inzoka (snake) and inyenzi (cockroach). It was the same kind of language that once was used jokingly, perhaps as a playful threat to kill Tutsi friends, even as they all socialized or worked together. In a mere four years, their toxic effect changed drastically and spread across the country like a virus, to refer to the entire Tutsi community. Not just nasty names, these words primed some Rwandan citizens to take action and kill those who were once their neighbors.

As Tirrell points out, when Donald Trump says “Syrian refugees are snakes,” he is likewise drawing on suggestive qualities about snakes that allow his listeners to infer certain negative ideas about an entire people, and by extension, other immigrants and refugees. Tirrell quotes Republican consultant Frank Luntz as saying that Trump initially left people “horrified by his offensive statements. But as time went on, they came to enjoy and absorb it.” As Tirrell elaborates, “Each repetition of an outrageous speech act makes the next one less of a surprise, until such speech becomes common enough to seem ‘normal,’ lowering the standard of acceptability.”

Toxic speech, under the insidious guise of ordinary language, can act as a virus if spread to the right crowd. It can be just as harmful as outright slurs, no longer the words we once knew.

You can read all our coverage of the coronavirus/COVID-19 outbreak for free. But please consider becoming a member on our Patreon page to support our nonprofit journalism.