“Falsehoods Fly: Why Misinformation Spreads and How to Stop It” by Paul Thagard. Columbia University Press

Full text

Vol.:(0123456789)
The Journal of Value Inquiry
https://doi.org/10.1007/s10790-024-09996-3
1 3
“Falsehoods Fly: Why Misinformation Spreads andHow
toStop It” byPaul Thagard. Columbia University Press
EricWinsberg1
Accepted: 10 June 2024
© The Author(s) 2024
Every so often, a book comes along that quickly proves that its main title sentence
is true. Falsehoods Fly is one such book. This book on “misinformation” is riddled
with claims that, by now, most reasonable people should see as false. This is par-
ticularly the case with regard to the book’s treatment of the epidemiology of SARS-
CoV-2, the evidence of its origins, and facts about the vaccines that mitigate some
of its harms. These are not just claims the author asserts to be true. They are claims
that the author thinks could only be denied or disbelieved by malicious or highly
motivated actors.
More on this soon. First, let’s back up. What is the book about and what are
its central claims? The central claim of the book is that “Misinformation cause[s]
deaths.” (p. 2) And that “Even when misinformation is not lethal, it can be
dangerous.”(p. 2) The book, we are told, “offers a systematic explanation of infor-
mation and misinformation, along with concrete advice on how improved thinking
and communication can benefit individuals and societies.” (p.2) But we should be
perfectly clear: while Thagard thinks that some of the problems of misinformation
can be remedied using the tools of critical thinking he offers, he is not nearly this
optimistic, nor are his suggestions anywhere this mild. Improved critical thinking
might be a salve for a select group of non-deplorables, Thagard argues.But for many
groups, “abolition is probably the best hope because changing their policies, val-
ues, and norms is unlikely to work.” He adds that “social media companies can be
lobbied to encourage them to limit the transmission of conspiracy theories more
strictly.”(p. 184).
It’s clear that, for Thagard, misinformation is a very urgent problem that war-
rants the curtailment of what is ordinarily regarded as protected speech.1 He notes
* Eric Winsberg
ew652@cam.ac.uk
1 University ofCambridge, Cambridge, UK
1 It’s clear that Thagard is arguing for the prohibition of speech (or the restriction of free assembly
on the basis of speech of those who assemble) that reaches far beyond what, for example, U.S. courts
regard as regulable speech. A canonical example of a claim he thinks warrants censoring is the claim he
calls the “replacement theory”: that low birth rates among European and North Americans of European
descent combined with high rates of immigration are part of an amorphous movement to “replace” Euro-
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

E. Winsberg
1 3
that “Since 2020, Twitter had a policy against coronavirus misinformation that
led to the suspension of more than eleven thousand accounts and removal of more
than 100,000 pieces of content.” (p.148) He laments that “This policy was ended
in November 2022.” (p.148) What he fails to note is that Mark Zuckerberg deeply
regrets much of what “the establishment” asked him to censor during the pandemic
because much of it turned out to be “debatable or true”2. Thagard also fails to note
that later reporting (based on information disclosed in lawsuits, data released after
Musk acquired Twitter/X, etc) revealed that many tweets that were suppressed by
Twitter contained nothing but CDC data!3
It seems, however, that this would not concern Thagard. After all, he approves of
a feature of a definition of misinformation4 that “includes information that is mis-
leading as well as false because even true reports can be harmful.” (p.3) Explain-
ing his view, he writes, “For example, someone who posts on social media that
a friend got a blood clot after getting a COVID-19 vaccination may be reporting
the truth, but the anecdote is misleading if it suggests that vaccines are dangerous
despite ample evidence that their risks are minuscule compared with the dangers of
COVID-19.”(p.3) In fact, I would argue that the quoted sentence (about miniscule
risks) is misleading, in the very sense Thagard makes so much of, for a number of
reasons. First,Thagard fails to note that at least one COVID vaccine (Johnson and
Johnson’s) was taken off the market precisely because of its risk of blood clots. Sec-
ond, what might matter is whether the risks are miniscule compared to the risk of
alternative vaccines. That’s in fact what the CDC decided. Third, whether the risks
of a vaccine are miniscule compared to the disease it is designed to mitigate depends
on how effectively it mitigates that risk, and how high the risk of the disease is
for the person in question. For a fifteen year old who has already been infected by
Covid-19, many experts believe the risks of various candidate vaccines outweigh the
degree of risk they mitigate.5 Indeed, the UK does not advise vaccinating children
under age 18, regardless of prior infection, and Germany and the Netherlands does
so only under special conditions 6 In the US, many children were de facto forced to
take these same vaccines, even after prior infection.
2 https:// lexfr idman. com/ mark- zucke rberg-2/#
3 https:// twitt er. com/ david zweig/ status/ 16073 87725 06523 2384
4 In the end, like most discussions of misinformation, he admits it is a concept that is incredibly difficult
to define. It requires “identifying the psychological and social mechanisms that produce information and
by describing how breakdowns in these mechanisms promote misinformation” (p.3).
5 https:// www. medpa getod ay. com/ opini on/ second- opini ons/ 93340
6 https:// www. medpa getod ay. com/ opini on/ second- opini ons/ 93340. (Here, the risk in question is myo-
carditis, not blood clots, but the details are irrelevant. Facebook and Twitter also censored or shadow
banned posts about myocarditis, and there’s nothing in Thagard’s book to suggest he doesn’t heartily
endorse this.)
peans with people of other ancestries. This claim is almost certainly false and obviously motivated by
pernicious racism, but doesn’t fit any of the obvious exceptions to freedom of speech, like defamation,
fighting words, speech integral to illegal conduct, speech that incites imminent lawless action, speech
that violates intellectual property law, or true threats. He also wants the government to coordinate with
social media companies to censor posts that are true but “misleading”.
Footnote 1 (Continued)
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1 3
“Falsehoods Fly: Why Misinformation Spreads and How to Stop It” by Paul Thagard. Columbia University Press
We needn’t adjudicate who is right on any of these topics to be worried about a
policy wherein the government pressures social media companies to censor posts
with documented truths in them because they might mislead someone into believ-
ing things that public health experts in the UK, Netherlands, and Germany believe.
This is especially true given that we know that the mechanisms being used at Twitter
(under the policy Thagard applauds and laments the ending of) to determine which
posts were “misleading” often involved either bots, or untrained low wage contrac-
tors in outsourcing facilities in places like the Philippines.7 It’s genuinely odd that
Thagard seems to think that bots and outsourcing facilities can identify misinfor-
mation, given that he thinks identifying it requires not only omniscience about the
first order facts, but being able to predict what truths will be misleading, and about
“the psychological and social mechanisms that produce information” and about how
breakdowns in these mechanisms promote” (p.3) what is shared.
Enough about what Thagard wants “abolished” or what he thinks social media
companies should be pressured by the government to censor, or even whom, as a
result of the misinformation they consume, “military forces may be required to con-
trol” (p.184). Most of these remarks are embarrassingly illiberal for a philosopher.
Sadly, I should remark, Thagard hardly needs singling out here. It is, in my opinion,
an embarrassment to our profession that a climate even exists where a philosopher
would express views like this, and expect (correctly!) to find a sympathetic audience
among other philosophers and a prestigious university press willing to publish it.8
What we should be discussing is not whether this part of the book is right, but what
made the disciplinary conditions conducive to the very possibility of a book like this
being published in philosophy. So let’s move on to his proposals to combat misinfor-
mation in a more liberal fashion: via “the process of correcting misinformation by
repairing or remedying it to restore, reclaim, or recover real information. The goal
is to explain how information often works well, sometimes breaks into misinforma-
tion, but can be mended by reinformation.”
The proposal is actually well summarized in the following table. (p.8)
TABLE 1.2 Profiles of real information and misinformation
Process Real information Misinformation
Acquisition Collecting by perception, instruments,
systematic observations, and controlled
experiments
Making stuff up, faulty observation, sloppy
experiments
Inference Evidence-based casual reasoning Motivated reasoning, flawed casual reasoning
Memory Evaluation-based storing and retrieving Motivated storing and retrieving
Spread Evaluation-based sending and receiving Motivated sending and receiving
He calls this “AIMS”. The idea is that responsible knowers Acquire, Infer,
reMember, and Spread information in completely different ways than misinformation
7 https:// twitt er. com/ david zweig/ status/ 16073 85019 10188 4418
8 Let me clear that I am not criticizing the press for publishing the book. I am criticizing the community
for producing a climate in which the press correctly perceived an appetite for this book.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

E. Winsberg
1 3
mongers. Of course we can all agree that it is better to acquire information using
perception, instruments, systematic observations, and controlled experiments, than it
is to “make stuff up”. And we can all agree that motivated reasoning is bad, and that
it is better to follow the evidence where it leads you than to jump to the conclusions
that one likes. The problems that I have with what Thagard is up to here, however,
are twofold.
First, he has an unbelievably naïve, simplistic, and I would even say patronizing
view of where the claims he disagrees with come from. He seems to think that it’s a
very simple matter to determine which of a large set of claims are the well supported
ones and which ones are peddled by mindless simpletons who don’t follow any basic
epistemic norms. He thinks, in short, that it’s very easy to tell what’s true and what’s
false about almost all the important topics of the day. But as I will show, when it
comes to his discussion of topics surrounding the pandemic, almost everything he
says is either demonstrably false, or, I would at least argue, not the most reason-
able thing to believe given present evidence. This is especially ironic given that he
reports that it was the spread of misinformation during the pandemic that motivated
him to write the book. I’ll have nothing to say about what he says about the war in
Ukraine or about political conspiracy theories. (I’m going to especially avoid the
topic of Ukraine because, other than to simply state that, like all unprovoked aggres-
sion, the Russian invasion was a morally heinous war crime, I personally think it’s
just too epistemically risky for me to comment on the details of a war as it pro-
gresses.) I’ll have a little bit to say about his chapters on the science of climate
change and on inequality research, but mostly to make my second point. My second
point is that even when he correctly identifies instances in which what appear to be
factual disagreements are, in actuality, disagreements about values, he has some odd
beliefs about how values operate and when they can be objectively adjudicated.
Let’s start with the pandemic. I will focus on four sets of claims Thagard makes
that I regard to be completely factually wrong. I do this in part to set the record
straight (since, after all, falsehoods fly), and in part to illustrate that determining
what is true on many important topics of the day is hard. I doubt Thagard believes
he violated his AIMS when he formed his beliefs about the pandemic, and yet, I’m
here to argue that he has formed poorly evidenced beliefs about a number of impor-
tant matters. I also hope that readers will be persuaded, inter alia, that I came to have
true beliefs about the pandemic, in part, by having access to the speech of people
who, if Thagard had had his way, would have been silenced.
The four categories of claims have to do with
• Covid-19 origins
• the epidemiological models that guided pandemic mitigation policy in many
places
• the efficacy of mask mandates and public mask-wearing recommendations
• the efficacy of Covid-19 vaccines in preventing community spread, as well as
their risk-benefit profile for select members of the public (a topic which we
already covered above.)
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1 3
“Falsehoods Fly: Why Misinformation Spreads and How to Stop It” by Paul Thagard. Columbia University Press
Let’s start with the origins of the virus that causes Covid-19. It’s quite clear that
Thagard regards the hypothesis (call it “lab leak”) that the Wuhan Institute of Virol-
ogy was “conducting research on a collected or manipulated virus that infected lab
researchers and then spread to the general population in Wuhan and the rest of the
world” (p.91) to be misinformation. He says “The problem with this hypothesis is
that the intermediate links in the causal chain that it proposes have not been substan-
tiated. No evidence shows that the Wuhan institute was manipulating or studying
the novel coronavirus…” (p. 91) He goes on to say that, because there were incom-
plete causal chains in the zoonosis hypothesis as well, there was a time when it was
appropriate to “refrain from making any inference until the gaps in at least one of
the explanations [were] filled in.”(p. 91) But this situation changed, he says, in 2022
when a paper published in Science (usually known as Worobey etal9) showed that
“The earliest known cases of COVID-19 in December 2019 were geographically
centered on the Wuhan market that sold several species of animals susceptible to
coronaviruses. This centering is easily explained by spread of the disease through
the animals, but the alternative lab-escape theory has to make additional assump-
tions about the virus spreading from the lab to the market.” (p.91, Thagard’s para-
phrase). Subsequent to this, according to Thagard, belief in lab leak could only be
explained by “motivated reasoning to blame China for the disease.” (p.91)10
Let’s start with the claim that “no evidence shows that the Wuhan institute was
manipulating” the novel virus. This is about as close as one can get to a documented
falsehood. Since Thagard holds Worobey etal in high regard, let’s consider what one
of its authors, Kristian Anderson, thinks about this claim. Andersen, by the way, is
the virologist who, submitted a paper on Feb 4th 2020 called “The proximal origins
of SARS-CoV-2”11that concluded “Our analyses clearly show that SARS-CoV-2 is
not a laboratory construct or a purposefully manipulated virus,” but just three days
earlier had privately shared with his co-authors that “I think the main thing still in
my mind is that the lab escape version of this is so friggin’ likely to have happened
because they were already doing this type of work and the molecular data is fully
consistent with that scenario.”12 Here is what Sir Jeremy Farrar, Chief Scientist at
the World Health Organization, said about Andersen’s reaction to his first discov-
ery of the earliest evidence that the WIV was in fact manipulating the virus. “Then
Kristian delivered his denouement: he’d found a scientific paper where exactly this
technique had been used to modify the spike protein of the original SARS-CoV-1
virus. At first glance, the paper Kristian had unearthed looked like a how-to manual
for building the Wuhan coronavirus in a laboratory. [Andersen and “Proximal Ori-
gins” co-author Edward Holmes] knew of a laboratory where researchers had been
experimenting on coronaviruses for years: the Wuhan Institute of Virology”. “‘Fuck,
9 https:// www. scien ce. org/ doi/ 10. 1126/ scien ce. abp87 15
10 I’ll note for the record that my credence in the lab leak hypothesis is very high, but certainly not
because of motivated reasoning. That this is so if evident from the fact that it went from very low to very
high when new evidence came out in roughly early 2022.
11 https:// www. nature. com/ artic les/ s41591- 020- 0820-9
12 https:// usrtk. org/ wp- conte nt/ uploa ds/ 2023/ 11/ so- frigg in- likely. png
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

E. Winsberg
1 3
this is bad’ was [Holmes]’s reaction to Kristian’s observation. His second instinct
was to call me on the burner phone.”13 (p.47)
This much has been known since 2020. But we have learned much more since:
Freedom of Information Act requests have unearthed a 2018 grant proposal to
DARPA called “Project Defuse: Defusing the threat of bat-borne coronaviruses.”
These documents contain a wealth of evidence that SARS-CoV-2 was manipu-
lated at the WIV. This includes not only evidence of plans to do something exactly
like building SARS-CoV-2 at WIV, but also molecular evidence that the virus is
manipulated.14
Is any of the above proof that the virus was manipulated at the WIV? Of course
not. Unless China hands over all of the laboratory’s records (something they have
not done and will almost certainly never do) we will never have proof. Is it remotely
consistent with the claim that “no evidence shows that the Wuhan institute was
manipulating” the novel virus? Not on any theory of evidence that I am familiar
with.
What about the claim that Worobey et al. tipped the scales (balancing lab leak
with zoonosis) so far that only motivated reasoning, at this point, explains anyone’s
belief in lab leak? Recall, first of all, that Worobey etal. was also co-authored by
Andersen and Holmes, the very people who, in February 2020, submitted a paper
to Nature Medicine that said “Our analyses clearly show that SARS-CoV-2 is not
a laboratory construct or a purposefully manipulated virus” despite having said,
over the previous three days that lab leak was “so friggin’ likely” and “Fuck, this
is bad”, right before calling Sir Jeremy Farrar “on the burner phone.” Much more
importantly, Worobey etal. is a deeply flawed paper. There are two problems with
this paper. The first is the model, the second is the data the model relies on. We’ll
start with the data. First of all, we know that the number of identified cases is much
smaller than the total number of people who would have been infected. The authors
of “The Huanan Seafood …” insist that that the data they have is representative.
They deny, in other words, that there was “ascertainment” bias in collecting the
cases. Ascertainment bias is what you would have if you looked for discarded bot-
tlecaps in the dark, and concluded that they were primarily discarded in the vicinity
of lamp posts. But recently, George Gao, director of the Chinese Center for Disease
Control, admitted that there almost certainly was ascertainment bias in the collec-
tion of cases. He says there was special focus on the Seafood Market and people
associated with it in the early search for cases.15 It has also been shows that someone
deleted some of the early sequence data from the NIH’s “Sequence Read Archive”.16
Some of this data has been recovered and suggests that there were cases with a
closer ancestry to bat viruses than any of the Seafood Market cases.17
15 https:// www. bbc. co. uk/ sounds/ play/ m001n g7c (Begin listening around 24:30 if you want to skip to the
relevant comment.) Alina Chan has highlighted this fact.
16 https:// www. washi ngton post. com/ opini ons/ 2022/ 11/ 17/ covid- early- cases- wuhan- china- myste ry/
17 https:// www. biorx iv. org/ conte nt/ 10. 1101/ 2021. 06. 18. 44905 1v1
13 https:// books. google. ca/ books/ about/ Spike. html? id= 7JkfE AAAQB AJ& redir_ esc=y
14 See https:// ericw insbe rg. subst ack. com/p/ oh- my- god- theres- been- an- outbr eak for many more details.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1 3
“Falsehoods Fly: Why Misinformation Spreads and How to Stop It” by Paul Thagard. Columbia University Press
What about the model? In January 2024, Stoyan and Chiu (2024), published “Sta-
tistics did not prove that the Huanan Seafood Wholesale Market was the early epi-
centre of the COVID-19 pandemic.” in the Journal of the Royal Statistical Society
Series A.18 Even taking the data at face value, they argued that “statistical conclu-
sion is invalid on two grounds: (a) The assumption that a centroid of early case loca-
tions or another simply constructed point is the origin of an epidemic is unproved.
(b) A Monte Carlo test used to conclude that no other location than the seafood mar-
ket can be the origin is flawed. Hence, the question of the origin of the pandemic has
not been answered by their statistical analysis.”
Perhaps more alarmingly, in a subsequent blog post on the website of the Institute
of Mathematical Statistics, they wrote: “We are pleased to have published our cri-
tique. However, we are astonished that no other colleagues have reported similar
findings. This situation raises doubts about the current state of the system of mod-
ern science and the general understanding of basic principles of statistics in modern
society, not to speak about the reviewing process of Science.”19
Let me be perfectly clear what I think is going on here. “Lab leak” was labelled
“misinformation” in a very successful propaganda campaign by people who were,
for reasons that aren’t fully clear, terrified enough that the hypothesis would gain
traction that they communicated about it over a burner phone. (As we will see later,
this is just how people who were terrified of a Donald Trump re-election, and peo-
ple terrified that people would decide they didn’t support lockdowns, behaved.) Peo-
ple like Paul Thagard, the author of this book, as well as other philosophers, and
many other academics, have convinced people that misinformation is a real thing
and poses a real danger. And these two things in combination have corrupted the
“current state of the system of modern science” and “the reviewing process of Sci-
ence.” It has also, as we will see, corrupted journalism and the proper role that the
state plays in regulating protected speech. To the philosophers and other academics
engaged in these discussions of misinformation: would youpleasejust stop?
Let’s turn next to a claim that Thagard makes about Covid-19 vaccines. He says
that clinical trials have shown that “vaccines, produced by Pfizer, Moderna, Astra-
Zeneca, and other companies, are effective at preventing the spread …of COVID-
19.” This is false. In fact, in October of 2022 “Janine Small, president of interna-
tional markets at Pfizer, told the European Parliament on Monday that Pfizer did
not know whether its COVID-19 vaccine prevented transmission of the virus before
it entered the market in December 2020.”20 In keeping with the kind of strategies
Thagard champions, every major media outlet labelled the retransmission of this
fact “misleading” because (here quoting the Associated Press fact checkers) “Pfizer
never claimed to have studied the issue before the vaccine’s market release.”21 The
Moderna trial tested for asymptomatic infection while Pfizer did not. No clinical
18 https:// www. biorx iv. org/ conte nt/ 10. 1101/ 2021. 06. 18. 44905 1v1
19 https:// imstat. org/ 2024/ 02/ 15/ chall enging- the- misuse- of- stati stics- in- the- quest ion- of- the- origin- of-
covid- 19/
20 https:// apnews. com/ artic le/ fact- check- pfizer- trans missi on- europ ean- parli ament- 95041 38632 26
21 https:// apnews. com/ artic le/ fact- check- pfizer- trans missi on- europ ean- parli ament- 95041 38632 26
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

E. Winsberg
1 3
trial tested for transmission.22 Did subsequent clinical trials establish this? No. We
will never know whether the original vaccines would have prevented the spread of
the early strains of the virus because this was never studied, and by the time the
vaccines were released, the early strains were displaced. During Omicron ,a Dan-
ish household study found no difference in transmission rates based on vaccination
status.23 Even for the booster, we can see that the cumulative infection rate ends
up being the same after around 160 days.24 Despite this, vaccine mandates were
imposed all over the world, with justifications offered in terms of their public health
benefits. This was almost certainly facilitated by people who labelled claims that
‘the vaccines don’t prevent transmission’ as misinformation.
Finally, let’s consider Thagard’s claim that: “A group led by Neil Ferguson at
Imperial College London began in January 2020 to use a variety of mathematical
techniques to build computer models to predict the course of the pandemic and the
effects of different strategies for dealing with it. These models unavoidably make
assumptions such as the reproduction rate of viral infections and the proportion of
infected people who will die, and risk the danger of the existence of unknown fac-
tors that would invalidate the model’s predictions. Nevertheless, the models are fre-
quently revised to account for the most recent data about the pandemic. “ (p. 90, my
emphasis)
When Ferguson et al. published the famous “Report 9”25 in March 2020, the
fatality rates and hospitalization rates per age decade in the model came from a
paper known as "Verity et al."26 which used the number of deaths in Hubei prov-
ince reports (the numerator) from newspaper and estimated the rate of infection (the
denominator) by looking at the repatriation flights of non-Chinese citizens in Wuhan
back to Europe and the US. Six people, in total, tested positive on those flights, and
so that n=6 was used to estimate the fatality rate of the virus in the ICL model. This
is consistent with Thagard’s claim that they “unavoidedly made assumptions”27. To
be clear, these assumptions were far from accurate. Verity etal used an infection
fatality rate (IFR) for people between the ages of 20-29 of 0.03%. Contemporary
estimates of the IFR in that age bracket (in the absence of vaccination or prior infec-
tion) are about 15 times lower than that (0.002%)28
But what about Thagard’s claim that the models were frequently revised to
account for the most recent data? The ICL group never did a sensitivity analysis on
23 https:// www. nature. com/ artic les/ s41467- 022- 33328-3
24 https:// www. nejm. org/ doi/ full/ 10. 1056/ NEJMo a2210 058. See especially figure1.
25 https:// www. imper ial. ac. uk/ mrc- global- infec tious- disea se- analy sis/ disea se- areas/ covid- 19/ report- 9-
impact- of- npis- on- covid- 19/
26 https:// www. thela ncet. com/ artic le/ S1473- 3099(20) 30243-7/ fullt ext
27 Although it wasn’t completely unavoidable, as there was by that time the data that came from the Dia-
mond Princess data. Verity at all knew of these data, and mentioned them, but threw the away. In retro-
spect they would have provided by estimates. And prospectively, an easy argument could have been made
that they were better, as they involved no estimation whatsoever of either denominators or numerators.
28 https:// www. ncbi. nlm. nih. gov/ pmc/ artic les/ PMC96 13797/
22 It’s a tricky affair to document a negative claim like this, but I have available, for anyone who doubts
this claim, confirmation in an email from Tracy Høeg, Dept of Epidemiology & Biostatistics- University
of California-San Francisco. I thank her for this as well as for providing the following two references.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1 3
“Falsehoods Fly: Why Misinformation Spreads and How to Stop It” by Paul Thagard. Columbia University Press
the parameters in the their model. But in June, 2020, sensitivity analyses began to
show that the ICL model “lacks practical parameter identifiability from data. The
analysis also showed that this limitation is fundamental, and not something read-
ily resolved should the model be driven with data of higher reliability.”29 A more
detailed study came out in November, 2020.30 It found that almost two-thirds of the
differences in the model’s results (with respect to how many lives suppression could
save) could be attributed to changes in just three especially important variables. It
was unable to closely reproduce the real data for any settings in the model with any
parameter values. And it showed that for most values of the three parameters, five to
six times as many people die during “maximum suppression” than the model pre-
dicted using the values that the ICL grouped used. This was a clear sign that the
model might very well have large errors that were cancelling out, making it funda-
mentally useless for evaluating counterfactuals: overly high mortality of the virus,
failure to represent the mechanisms that make Covid-19 come in waves, and overly
high effectiveness of non-pharmaceutical interventions (NPIs). This suspicion was
greatly heightened by looking at what was happening in Sweden.
On March 26th, a few days after they published report 9, the ICL group published
Report 12.31 In it, they predicted that Sweden would have 90,157 deaths if they
did not go into maximum suppression. (Which they did not. Technically, Sweden
implemented only one of the many policies that constituted “suppression” in ICL-
speak: closing large events—in Sweden this was defined as events of over 50 peo-
ple.) According to the model, those deaths should have all happened by the end of
July, and the epidemic should have burned itself out. Of course, nothing like this
happened. According to Thagard, therefore, we should fully expect that by the end
of June, the ICL group would have radically recanted, and fundamentally altered
their model to show that lockdowns were much less effective and the virus was both
considerably less lethal, and much more susceptible to downturns in the absence of
lockdowns.
They did the opposite. The ICL group doubled down on their model. In June
2020, they published the infamous paper known as “Flaxman etal” (with Neil Fer-
guson as co-author).32 In it, they “showed” that lockdowns had saved three million
lives in Europe in the spring of 2020 (in a period ending May 4th). Shockingly, they
“showed” that Sweden had averted 26,000 deaths. With NPIs! How is this possible?
The model also “showed” that nothing other than NPIs was effective. The only NPI
Sweden employed was closing large events. The answer is that they made the effect
of closing large events in Sweden 45 times larger than it was in any other European
country. In most countries, the country-specific effect, for closing large events,on
the reproduction number of the virus was nearly zero. “Because of this country-spe-
cific effect, their model found that banning public events had reduced transmission
by ~72.2% in Sweden, but only by ~1.6% everywhere else. Moreover, according to
31 https:// www. imper ial. ac. uk/ mrc- global- infec tious- disea se- analy sis/ disea se- areas/ covid- 19/ report- 12-
global- impact- covid- 19/
32 https:// www. nature. com/ artic les/ s41586- 020- 2405-7
29 https:// www. medrx iv. org/ conte nt/ 10. 1101/ 2020. 06. 10. 20127 324v1
30 https:// www. nature. com/ artic les/ s43588- 021- 00028-9
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

E. Winsberg
1 3
the prior they used for the country-specific effect, the probability that it would be
that large was only ~0.025%.”33 Not only this, but none of these facts were revealed
in the printed paper, published in Nature. I’ll leave it as an exercise for the reader
to determine whether this confirms or refutes Thagard’s claim that “models are fre-
quently revised to account for the most recent data about the pandemic. Indeed, it
is well documented that the Covid-19 projection models continued to make rhetori-
cally powerful but wildly overly pessimistic projections in the absence of NPIs for
most of the pandemic34 and that very few of them did an adequate job of assessing
their own performance and adhering to best practices in model self-evaluation.35
The last thing I want to do is exasperate the reader with the most annoying debate
in the history of culture wars: the debate about the efficacy of masks and mask
mandates. Suffice it to say that Thagard is deeply confused about the evidence. His
diagram on page 89 clearly suggests he thinks that the 2023 Cochrane Report “Do
physical measures such as hand-washing or wearing masks stop or slow down the
spread of respiratory viruses?” is just one study that ought to be weighed against
other studies, such as the Bangladesh Mask study. But of course the Cochrane
group, who are world-leading experts in evidence-based medicine, weighed all the
available evidence themselves, obviously including the Bangladesh study36, and
concluded that “Wearing masks in the community probably makes little or no differ-
ence to the outcome of laboratory‐confirmed influenza/SARS‐CoV‐2 compared to
not wearing masks…The use of a N95/P2 respirators compared to medical/surgical
masks probably makes little or no difference for the objective and more precise out-
come of laboratory‐confirmed influenza infection…One previously reported ongo-
ing RCT has now been published and observed that medical/surgical masks were
non‐inferior to N95 respirators in a large study of 1009 healthcare workers in four
countries providing direct care to COVID‐19 patients.”
One reason that, pace Thagard, the Bangladesh Mask study failed to contribute
to a clear difference with the use of surgical masks (which the study employed)
is that the effect in that study was miniscule. That study found that among more
than 300,000 subjects allocated to a mask group and a non-mask group, the non-
mask group had 20 more cases of COVID over 6 months. What I don’t believe the
Cochrane group took into account was that even this tiny effect was almost certainly
due to ascertainment bias. As Chikina etal (2022) note: “Upon reanalysis, we find a
large, statistically significant imbalance in the size of the treatment and control arms
evincing substantial post-randomization ascertainment bias by unblinded staff. The
33 https:// necpl uribu simpar. net/ reply- to- andrew- gelman- and- flaxm an- et- al- on- the- effec tiven ess- of- non-
pharm aceut ical- inter venti ons/
34 https:// www. news- medic al. net/ news/ 20211 005/ How- accur ate- is- COVID- 19- model ling. aspx
35 https:// www. thela ncet. com/ journ als/ landig/ artic le/ PIIS2 589- 7500(22) 00148-0/ fullt ex
36 We included 11new RCTs and cluster‐RCTs (610,872 participants) in this update, bringing the total
number of RCTs to 78. Six of the new trials were conducted during the COVID‐19 pandemic; two from
Mexico, and one each from Denmark, Bangladesh, England, and Norway. We identified four ongoing
studies, of which one is completed, but unreported, evaluating masks concurrent with the COVID‐19
pandemic. https:// www. cochr aneli brary. com/ cdsr/ doi/ 10. 1002/ 14651 858. CD006 207. pub6/ full
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1 3
“Falsehoods Fly: Why Misinformation Spreads and How to Stop It” by Paul Thagard. Columbia University Press
observed decrease in the primary outcome is the same magnitude as the population
imbalance but fails significance by the same tests.”37
Enough about masks. Enough about all the things Paul Thagard labelled as mis-
information based on false reporting of the underlying facts. I’ll come back in a
moment to why I have subjected the reader to this tedious “fact-checking” of Thag-
ard’s book. Before that, I promised to say something about his chapter on Climate
Science and to elaborate on my claim that even when Thagard correctly identifies
instances in which what appear to be factual disagreements are, in actuality, disa-
greements about values, he has some odd beliefs about how values operate and when
they can be objectively adjudicated. Regarding people who are anti-egalitarian, he
summarizes their values as “freedom is a more fundamental value than equality.”
(p. 199) Let’s set aside the question about whether this is a correct diagnosis of the
values held by anti-egalitarians. (Many of them believe, I think, that overall prosper-
ity is more important than equality, rather than that freedom is. They believe free
markets lead to greater prosperity.) The more puzzling claim that Thagard makes
is that these claims can count as misinformation because some of these values are
objectively wrong. He writes:
“We can grant that values are emotional attitudes but still see them as poten-
tially objective given a rich theory of emotions. Emotions are not just bodily reac-
tion but also require cognitive appraisals of the significance of a situation for a per-
son’s goals. For example, my reaction to the prospect that 50 million people will
die because of climate change is visceral but it is also cognitive because my goals
include the flourishing of the human species.” (p.128)
The first thing I want to note is that this argument is odd, especially from a phi-
losopher. I don’t want to dispute that some values are objectively better than other
values. This is contentious, to be sure, but also surely something that many reason-
able people hold. But the argument given for that claim here is odd. Of course, if
my goal is to consume as much chocolate ice cream as possible, then valuing an
opportunity to eat vanilla ice cream over an opportunity to eat chocolate ice cream
is wrong. But this only shows that goals and values are intricately related. It doesn’t
show that valuing vanilla ice cream is objectively wrong. Maybe this worry of mine
is pedantic, but I have a deeper one, and it relates to my overall evaluation of his
chapter on climate change.
Unlike in the pandemic chapter, I agree that the claims Thagard labels misinfor-
mation are very, or even extremely, likely to be false. They are claims that, in my
opinion, it is extremely difficult for a reasonable person to affirm given the pres-
ently available evidence. The claims he subjects to his AIMS analysis in this chapter
are primarily limited to these three: that global warming is increasing, that human
influence is the main cause of this increase, and if humans continue to emit more
CO2 and other heat-trapping gases, the warming trend will continue into the future.
I’m hesitant to label the denial of these claims misinformation38, but only because
38 I’m happy to label some instances of the denial of these claims “disinformation” if it can be shown
that, in these instances, the actors in question wilfully misrepresented what they believed for financial
gain or other nefarious motives. This is a separate question.
37 https:// www. ncbi. nlm. nih. gov/ pmc/ artic les/ PMC94 79361/
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

E. Winsberg
1 3
I don’t love the term, not because I disagree that the people who deny them are
unreasonable.
But Thagard doesn’t stop here. He admits that “Values are an indispensable part
of decision making,” but goes on to say “applying them has two dangerous pitfalls.
First, good decisions depend on values that are legitimately based on universal
human needs. The values deployed by the IPCC…include avoiding harm, saving
lives, and promoting equality, which help to satisfy the needs of the huge majority
of people. In contrast, the values of climate change deniers are skewed toward the
needs of a small minority of rich and powerful people.”(p.136).
First of all, let’s back up one step. The use of the word “climate change deniers”
in that paragraph is a non-sequitur. The paragraph is about values and decision mak-
ing, not about fact evaluation. So, the relevant group of people here are not climate
change deniers, but opponents of various mitigation measures. More importantly,
Thagard seems absolutely convinced that the only reason you could oppose climate
mitigation is because you value “the needs of a small minority of a small minority of
rich and powerful people” over saving the lives of 50 million people. But this is con-
fused for a number of reasons. First of all, the claim that unmitigated climate change
will, in the foreseeable future, kill 50 million people is nowhere near as unconten-
tious as the negations of the claims, canvassed above, that Thagard argues are misin-
formation in the chapter called “storms”39.
People have various reasons for doubting this. Some believe climate repair strat-
egies can save us (a highly contentious claim but one that would be hard to label
misinformation. Indeed, it’s a claim that has recently been gaining traction in the
climate science community in the last few years.) Some believe that adaptation can
dramatically lower the harms of climate change. Some believe that the specific dam-
age that climate change will cause is too local for our best models to reliably project.
The important point is that you can’t have your cake and eat it here. You can’t limit
yourself to claims like “climate change is real” when you argue for about misinfor-
mation but then use claims like “climate change will kill 50 million people” when
you argue that some people are employing values that are not legitimately based on
human needs. Second, Thagard is going way out over his skis when he suggests that
those who oppose mitigation only value “the needs of a small minority of rich and
powerful people.” Consider, for example, that I vigorously advocate for the miti-
gation of climate change by aggressively building nuclear power plants around the
world. The many, many people who disagree with me, and therefore oppose an (in
my opinion) important strategy for mitigation, do not only value “the needs of a
small minority of rich and powerful people.” Many people, moreover, who oppose
other mitigation strategies do so because they fear that such strategies will slow eco-
nomic growth, and that this will disproportionately hurt the global poor, potentially
39 It’s slightly ironic that Thagard chose this title for the chapter, since, contrary to much popular belief,
there is very little consensus in the climate science community about the effect of climate change on
storms. Perhaps more precisely, there is overwhelming consensus that storms will be wetter, but much
uncertainty about whether they will change in frequency and intensity. And we do not have enough good
data from the past to say with any confidence whether storms are more frequent or more intense now
than they were, e.g., in the early 20th Century. https:// www. gfdl. noaa. gov/ global- warmi ng- and- hurri
canes/
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1 3
“Falsehoods Fly: Why Misinformation Spreads and How to Stop It” by Paul Thagard. Columbia University Press
leading to more deaths by economic deprivation than they save. Or they think that
politicians cannot be trusted to enact effective climate mitigation policies that won’t
do more harm than good. Or they think the political backlash from policies that
hurt rural people more than urban dwellers will be dangerous and bad. Again, one
doesn’t have to adjudicate any of these debates to see that it is unhelpful to accuse
such people of misinformation and having illegitimate values and lacking a rich set
of emotions. These are hard policy questions and the people who raise these objec-
tions deserve detailed, nuanced answers to their worries. It is not helpful to lump
them together with people who do not believe carbon-dioxide can warm the planet
or with people who think further enriching Jeffrey Bezos is more important than
saving lives (if any of the latter such people even exist).
Let me finally come back to why I have subjected the reader to the tedious “fact-
checking” of Thagard’s book. I believe “misinformation” accusations are a danger-
ous and illiberal propaganda tactic—nothing but a raw exercise of power—and I
want philosophers and other academics to understand how difficult, complex, and
sometimes tedious it can be to really get to the bottom of some of the factual dis-
putes that underly some of the biggest policy disputes of the day. This was true dur-
ing the pandemic, and, when it comes to policy-relevant claims, it’s true about cli-
mate science. It’s also true in politics. Recall the infamous story, which broke in
the New York Post in the leadup to the 2020 election40, detailing emails allegedly
found on the Biden son’s laptop. Accusations of misinformation flew (see what I
did there?) around almost immediately. Facebook blocked or restricted posts of the
newspaper story. Twitter completely banned them41, even deleting links to the story
in people’s personal direct messages42. Glenn Greenwald, a co-founder of The Inter-
cept, left the publication because they refused to publish his reporting on the story.
But most of the stories and allegations underlying the claim that the Post article
counted as misinformation have not held up as the owl of Minerva has taken flight.
In particular, the claim that the laptop itself was a Russian plant have been falsified,
and there is no disagreement among major sources that the laptop is real, that it
belonged to Hunter Biden, and that it made its way from Hunter Biden, to the com-
puter store, and then to the New York Post. And the New York Times reports that
many if not all the emails have been “authenticated”43.
Or consider another case more closely related to Thagard’s chapter on Covid-19.
Early in the pandemic, there was a desire, among some people (myself included),
for vigorous debate about the wisdom of lockdowns. This was partly true because
lockdowns contradicted almost all of the existing science and wisdom about how
to act in a pandemic. Sunetra Gupta of the University of Oxford, Jay Bhattacha-
rya of Stanford University, and Martin Kulldorff of Harvard University wanted such
debate, and signed an open letter to that effect called the “Great Barrington Declara-
tion” (GBD)44. Privately, Francis Collins, the director of the NIH, emailed Anthony
40 https:// nypost. com/ 2020/ 10/ 14/ email- revea ls- how- hunter- biden- intro duced- ukrai nian- biz- man- to- dad/
41 https:// www. vox. com/ 22992 772/ hunter- biden- laptop
42 https:// techc runch. com/ 2020/ 10/ 16/ twitt er- new- york- post
43 https:// www. nytim es. com/ 2022/ 03/ 16/ us/ polit ics/ hunter- biden- tax- bill- inves tigat ion. html
44 https:// gbdec larat ion. org/
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

E. Winsberg
1 3
Fauci, who needs no introduction, that “This proposal from the three fringe epide-
miologists…seems to be getting a lot of attention – and even a co-signature from
Nobel prize-winner Mike Leavitt at Stanford. There needs to be a quick and dev-
astating take down of its premises. I don’t see anything like that on line yet—is it
underway?”45 The devastating takedown was underway. After the publication of the
Great Barrington Declaration, the “fringe epidemiologists” were censored on social
media, Google deboosted search results for the GBD, putting pieces critical of it
above the link to the actual declaration.46 Reddit removed links to the Declaration
from COVID-19 policy discussion forums.47 In February 2021, Facebook deleted
the Great Barrington Declaration page for a week.48 On March 18, 2021, Governor
Ron DeSantis of Florida held a roundtable discussion with some of the GBD sig-
natories. YouTube removed the video, claiming that it "contradicts the consensus
of local and global health authorities regarding the efficacy of masks to prevent the
spread of COVID-19."49 Was this all for the best? Even Francis Collins now seems
to think it wasn’t. In a panel discussion, held in mid-2023, called “A deplorable and
an elitist walk into a bar”50 Collins said, “if you’re a public health person and you’re
trying to make a decision you have this very narrow view of what the right decision
is and that is something that will save a life… So you attach infinite value to stop-
ping the disease and saving a life [and] you attach a zero value to whether this actu-
ally totally disrupts people’s lives, ruins the economy and has many kids kept out of
school in a way that they never right quite recover…This is a public health mindset
and I think a lot of us involved in trying to make those recommendations had that
mindset and that was really unfortunate. That’s another mistake we made.”
But this “mistake” was literally exactly the one Gupta, Bhattacharya, and Kull-
dorff were trying to draw attention to in the open letter that Collins labelled misin-
formation and kept people from having access to.
Now, it’s possible to believe that the ends justified the means. Especially in the
first case. Maybe labelling the Hunter Biden laptop story misinformation, keeping
the New York Post story about it off of social media, and stopping Glen Greenwald
from writing about it in the Intercept, kept Donald Trump from being re-elected.
Maybe this saved democracy. So maybe it was worth it. But if you think this, you
45 https:// www. wsj. com/ artic les/ fauci- colli ns- emails- great- barri ngton- decla ration- covid- pande mic- lockd
own- 11640 129116
46 Fraser Myers (2020) “Why Has Google Censored the Great Barrington Declaration?” Spiked Online.
October 12, 2020. https:// www. spiked- online. com/ 2020/ 10/ 12/ why- has- google- censo red- the- great- barri
ngton- decla ration/
47 Ethan Yang (2020) “Reddit’s Censorship of The Great Barrington Declaration” American Institute
for Economic Policy Research. Oct. 8, 2020. https:// www. aier. org/ artic le/ reddi ts- censo rship- of- the- great-
barri ngton decla ration/
48 Daniel Payne (2021) “Facebook removes page of international disease experts critical of COVID
lockdowns” Just the News. February 5, 2021. https:// justt henews. com/ nation/ techn ology/ faceb ook- remov
es- page- inter natio naldi sease- exper ts- who- have- been- criti cal-
49 4 Wall Street Journal Editorial Board. (2021) “YouTube’s Assault on Covid Accountability” Wall
Street Journal. April 8, 2021. https:// www. wsj. com/ artic les/ youtu bes- assau lt- on- covid- accou ntabi lity-
11617 921149
50 https:// youtu. be/ W1eAv h1sWiw? si= Y1DFO hX8vl N8PUux
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1 3
“Falsehoods Fly: Why Misinformation Spreads and How to Stop It” by Paul Thagard. Columbia University Press
should write a book entitled “Falsehoods Fly: Misinformation and how to use it for
political advantage.” I would have declined to review such a book.
Funding Funding was provided by British Academy (GP/400185).
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License,
which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long
as you give appropriate credit to the original author(s) and the source, provide a link to the Creative
Commons licence, and indicate if changes were made. The images or other third party material in
this article are included in the article’s Creative Commons licence, unless indicated otherwise in
a credit line to the material. If material is not included in the article’s Creative Commons licence and
your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need
to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://
creativecommons.org/licenses/by/4.0/.
Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps
and institutional affiliations.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center
GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers
and authorised users (“Users”), for small-scale personal, non-commercial use provided that all
copyright, trade and service marks and other proprietary notices are maintained. By accessing,
sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of
use (“Terms”). For these purposes, Springer Nature considers academic use (by researchers and
students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and
conditions, a relevant site licence or a personal subscription. These Terms will prevail over any
conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription (to
the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of
the Creative Commons license used will apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may
also use these personal data internally within ResearchGate and Springer Nature and as agreed share
it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not otherwise
disclose your personal data outside the ResearchGate or the Springer Nature group of companies
unless we have your permission as detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial
use, it is important to note that Users may not:
use such content for the purpose of providing other users with access on a regular or large scale
basis or as a means to circumvent access control;
use such content where to do so would be considered a criminal or statutory offence in any
jurisdiction, or gives rise to civil liability, or is otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association
unless explicitly agreed to by Springer Nature in writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a
systematic database of Springer Nature journal content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a
product or service that creates revenue, royalties, rent or income from our content or its inclusion as
part of a paid for service or for other commercial gain. Springer Nature journal content cannot be
used for inter-library loans and librarians may not upload Springer Nature journal content on a large
scale into their, or any other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not
obligated to publish any information or content on this website and may remove it or features or
functionality at our sole discretion, at any time with or without notice. Springer Nature may revoke
this licence to you at any time and remove access to any copies of the Springer Nature journal content
which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or
guarantees to Users, either express or implied with respect to the Springer nature journal content and
all parties disclaim and waive any implied warranties or warranties imposed by law, including
merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published
by Springer Nature that may be licensed from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a
regular basis or in any other manner not expressly permitted by these Terms, please contact Springer
Nature at
onlineservice@springernature.com