Investigator Calls For An "Epidemiology Of Toxic Speech"
Thought-Provoking Interview With The University of Connecticut’s Lynne
“Toxic speech, like
any toxin, is a threat to the well-being and even the very lives of
those against who it is deployed…Thinking in epidemiological terms
highlights that toxic speech is a community problem in need of social
These are among the points made by the University of
in a paper published in 2017 entitled “Toxic Speech: Toward an
Epidemiology of Discursive Harm.” In her paper, Tirrell calls for an
“epidemiology of discursive toxicity”.
The Epidemiology Monitor is always on the lookout for original ways of
using epidemiology to share these with our readers. We interviewed
Tirrell to learn more about her ideas and proposal. Whether you are
skeptical or not, read on for an eye-opening perspective about the
potential health impact of the words we use to talk to one another,
and why epidemiologists should get involved.
EM: How did you get
interested in the impact of toxic speech, what led you to this?
It is said that Gandhi held that a language is a reflection of the
health of a society. How we talk to and about each other speaks
volumes about us, and has a deep effect on our health and well-being.
In my research, I started out looking at the power of derogatory terms
(racist and sexist epithets) to inflict harm. Not just hurt someone’s
feelings, but truly harm them even when their feelings might not be
hurt at all. Later, research took me to Rwanda to learn about the role
of divisive discourse, especially derogatory terms for others, in
preparing ordinary citizens to participate in killing their neighbors.
I developed an account to explain the linguistic mechanisms involved.
That got me thinking about background conditions, and the ways that a
culture can be “seeded” with toxic concepts that can weaken some
people and then eventually kill them. Words alone didn’t kill the
Tutsi—their Hutu neighbors did that—but ongoing changing linguistic
practices were a necessary part of making it happen. Words are never
alone. That project led to thinking about how we might track those
changes, or how to assess the linguistic health of a society
or community. If we can spot the harms and the patterns of their
distribution, maybe we can prevent greater harms. I’m focused on
assessment of risk.
What is the toxic
speech you are interested in? Which one not?
I look at toxic speech practices, not one-off cruelties or
damaging remarks. These practices reveal the ways that people use
harmful language as a systemic form of control. Such speech may
promote discrimination or deprive people of important powers of
self-determination and social and civic participation. Racist, sexist,
homophobic slurs count as systemic toxic speech that generally serve
to lower the baseline well-being of the people targeted. This is now
well studied. It’s easy to identify the epithet that’s thrown
one-on-one; it’s like a punch in the nose. Certain speech practices
are more subtle and insidious, and I think their capacity to harm us
can be devastating and yet hard to prove.
It’s also important to
think about how the position of the speaker and the relationship
between speaker, hearer, or target, can enhance or diminish the power
of the speech act. Parents have more power to harm their child in
certain ways than a stranger, and peers have yet different powers over
the child. Tracking the relationships within which certain kinds of
harms are delivered would be informative, and might help us see ways
to foster changes. The results of such surveillance might be
How does toxic speech as a topic or subject differ from other concerns
of epidemiologists like disease, accidents, and other conditions?
Many harms are the same. Perhaps the mechanisms of delivery vary, but
the result is poor health, physical, cognitive, and emotional
dysfunction. Epidemiology could contribute to an empirical
understanding of the power of speech to help or harm people. I’ve
scoped out an overall argument that speech can be toxic, but the
empirical research would address these questions: Which populations
are harmed? How is their well-being damaged? What triggers the toxic
effect? Are there inoculating protections? Are there antidotes?
It’s my view that
toxic speech plays a significant role in a variety of physical and
mental health problems. My work so far has been urging philosophers
who think about the harmful power of language to investigate the
usefulness of toxicity for understanding harms that are delivered in
more subtle or diffuse ways. An epidemiology of toxic speech could
help us make clearer what counts as a “chilly climate” or progresses
to a toxic one.
If you start from the
harm and work backward, the harms of speech often get categorized in
some other way. Why is this child depressed and anxious? I’m saying
caregivers should examine the child’s expressive environment in
addition to the things they ordinarily consider. It would be easiest
to start with mental disorders, but many physical disorders could be
launching pads too, particularly those connected to stress.
EM: Is it more difficult to define the harms than other topics
epidemiologists work on?
Lots of public health issues may be like toxic speech
such as bullying, or poor parenting, or other suggested public health
issues. Presumably, making toxic speech a public health issue would
equate to getting epidemiologists involved since epidemiology is the
basic science of public health.
Let’s think about how speech is the key mechanism of bullying. No one
doubts that bullying is harmful, so we can study what bullies actually
say, as a mechanism for delivering harm. (Similarly, poor parenting
often involves barraging disparaging remarks upon the child.) In these
social problems, we need to attend to the power of the speech used,
how it effects those targeted, and the kinds of licenses it issues to
everyone who hears it. Those licenses are something I would like to
see tracked. We could track the increase of frequency of use, the
contexts in which it is used, for what purpose, and track the
outcomes. When is bullying effective at damaging, and when does the
target have the wherewithal to resist? What exactly is that
“wherewithal”? That’s just one explicit example. As the studies
continue, we can move to more subtle cases.
EM: So, what would
the epidemiology of toxic speech look like to you, or what info would
an epidemiologic profile or approach provide you with that is useful?
It would be important to look at the prevalence of toxic speech in a
population, the who, what, when, where, why of it. How frequent is
it, how ubiquitous is it, what kinds of damage does it inflict?
Suppose you have a community in which there is a spike in suicide
rates amongst teens. An epidemiologist could look at what those teens
are saying to and about each other, and track the frequency of
linguistic disparagement in all degrees. Not only the direct attacks,
but third-person uses that others hear and take up. And not only
negative messages, but glorifications of suicides, practices of urging
people to do it and so on. Are they sponsoring each other, through
speech? I would also look at what they are reading, watching, and
taking in from various sources: what’s their expressive diet?
Eliminating the toxins might not be possible or even desirable, but
strengthening their resistance or immunity might be well within reach
once you know what’s going on.
To assess civic
health, it would be worth tracking the kinds of derogatory terms
people use against and about each other, how much polarization those
uses bring about, to what degree it leads to individual and group
isolation, and then evaluate the health of the polarized groups. It
would be enlightening to get a handle on the frequency of occurrence
of these terms and how they are used.
In language, we look
at speaker, hearers—some of whom might be targets of the speech and
some might be bystanders. On my view, we look at the ways that saying
something can change what others think they have permission to say.
This fits with a viral conception—the form of speech spreads through
contact. One person slams another with a slur, and then someone
hearing it might pick it up and use it. Not everyone will take it up,
but still, the usage spreads, person to person, like a contagion.
EM: What preventive measures
can we imagine being practical to implement that would not violate
free speech? Opinions might differ on this point. Presumably we have
confronted these issues before by regulating movie content for
children, TV watching time, and so forth.
I don’t want
to advocate for a language-police. In the US, we have taken the first
amendment to apply well beyond its initial scope (only what the
government cannot restrict), and this is a double-edged sword. We do
need to protect expression, because that is also important to the
health and well-being of the person, but we also need to protect each
other from linguistic violence. Philosophers and jurists are working
on balancing these concerns. An epidemiologist could help by showing
where the balance tilts in one way or another.
If we learn more about how some forms of speech create
toxic environments, or how others deliver a heavy dose of harm at
once, then we can find ways to mitigate those harms. Susceptibility
surely varies across populations, so tracking resilience would be very
helpful. Where resilience is impossible, then restricting the speech
would seem well justified. This isn’t always going to be a matter of
law. It will often be about instituting new social norms. It isn’t a
law that parents wait 24 hours after their child is fever-free to
return to school. It is a demand of the public health departments of
most towns, and once parents know, they can comply or not. In my trips
to Rwanda, I learned ways to speak to avoid triggering the PTSD from
which many survivors suffer. Social norms do tremendous work to
protect the vulnerable.
What is the best evidence we can find for the negative effects of
toxic speech? Clear cut examples?
Wherever you see propaganda emerge, especially propaganda that targets
a segment of the population, it should be possible to track the
effects of that propaganda on both the in-group and the out-group. In
the US and across the EU we have seen increasing use of “terrorist” as
a synonym for “Muslim”, so what can we discover about the effects of
this on the well-being of Muslims in these regions? The easiest thing
would be to track direct physical attacks, clear cases of
discrimination, and so on, but there is more work to do. Those visible
signs are the tip of the ice-berg. For example, in a community with
high prevalence of derogatory language about Muslims, it would also be
important to look at the frequency of school truancy in Muslim youth,
health issues reported in the schools, job loss and job changes in
adults, and so on. These are measures of insecurity, which can lead to
physical and mental health issues. Maybe an epidemiologist could also
track the rate of access to health care in these areas, comparing
Muslims to non-Muslims.
EM: Can we put
together a short description of the case you are making---why an
enterprising young epidemiologist or other investigator might want to
undertake something like this?
If we want to promote the health and well-being of our society, we
need to promote both physical and mental health. Each of us is caught
up in complex identities, and these have different values to our
communities. The ways that we talk to and about each other actually
open and close possibilities in our social interactions, in our
capacity to live the lives we choose.
For other toxins,
epidemiologists track the incidence of harm, routes of exposure,
frequency, avoidability or inavoidability, and more. We should do the
same for toxic speech. Some toxic speech becomes woven into the fabric
of society; such cases are easier to see when social change occurs.
For another example,
the prevalence of pro-heterosexual and anti-homosexual discourse has
been part and parcel of the discrimination imposed on gay men and
lesbians. Casual comments and value-laden remarks, tossed off
matter-of-factly as socially accepted, served to keep 10% of the
population in the closet, hiding their desires, and often taking risks
that led to mental and physical harms. The CDC reports that that LGBT
teens have significantly higher rates of suicidal ideation (3x) and
much higher rates of suicide (5x) than their heterosexual
counterparts. American culture has been toxic for gay men and
lesbians, a toxicity delivered by speech, sometimes enforced by
violent actions. Whether this improves is something to track with
changes in law and social norms. The CDC tracks teen suicide rates,
has made the connection between bullying and LGBT suicide prevalence,
so my suggestion is to deepen the study to to examine the discourse
practices in the schools with high prevalence, for example. It would
be great to hear what an epidemiologist could envision tracking.
An epidemiologist who
includes the concept of toxic speech in his or her work and tracks the
speech practices associated with diseases they are tracking, may well
have a better
chance of nipping the
outbreak in the bud. Considering the teen suicide example, if an
epidemiologist could discern toxic speech practices in one outbreak,
and then the next, a pattern might emerge that might help us to keep
teens alive long enough to grow up to thrive. These patterns might
turn out to be very local, but broader patterns and lessons might
emerge over time.