Denialism as Applied Skepticism: Philosophical and Empirical Considerations

Abstract

The scientific community, we hold, often provides society with knowledge—that the HIV virus causes AIDS, that anthropogenic climate change is underway, that the MMR vaccine is safe. Some deny that we have this knowledge, however, and work to undermine it in others. It has been common (but not uncontroversial) to refer to such agents as “denialists”. At first glance, then, denialism appears to be a form of skepticism. But while we know that various denialist strategies for suppressing belief are generally effective, little is known about which strategies are most effective. We see this as an important first step toward their remediation. This paper leverages the approximate comparison to various forms of philosophical skepticism to design an experimental test of the efficacy of four broad strategies of denial at suppressing belief in specific scientific claims. Our results suggest that assertive strategies are more effective at suppressing belief than questioning strategies.

Full text

Denialism as Applied Skepticism:
Philosophical and Empirical Considerations
Matthew H. Slater
1
•Joanna K. Huxster
2
•Julia E. Bresticker
3
•
Victor LoPiccolo
4
Received: 11 January 2018 / Accepted: 10 August 2018
Springer Nature B.V. 2018
Abstract
The scientific community, we hold, often provides society with knowledge—that
the HIV virus causes AIDS, that anthropogenic climate change is underway, that the
MMR vaccine is safe. Some deny that we have this knowledge, however, and work
to undermine it in others. It has been common (but not uncontroversial) to refer to
such agents as ‘‘denialists’’. At first glance, then, denialism appears to be a form of
skepticism. But while we know that various denialist strategies for suppressing
belief are generally effective, little is known about which strategies are most
effective. We see this as an important first step toward their remediation. This paper
leverages the approximate comparison to various forms of philosophical skepticism
to design an experimental test of the efficacy of four broad strategies of denial at
suppressing belief in specific scientific claims. Our results suggest that assertive
strategies are more effective at suppressing belief than questioning strategies.
Electronic supplementary material The online version of this article (https://doi.org/10.1007/s10670-
018-0054-0) contains supplementary material, which is available to authorized users.
&Matthew H. Slater
matthew.slater@gmail.com
Joanna K. Huxster
jhuxster@gmail.com
Julia E. Bresticker
jeb073@bucknell.edu
Victor LoPiccolo
vjl003@bucknell.edu
1
Department of Philosophy, Bucknell University, Lewisburg, USA
2
Department of Environmental Studies, Eckerd College, Saint Petersburg, USA
3
Department of Physics, Bucknell University, Lewisburg, USA
4
Department of Education, Bucknell University, Lewisburg, USA
123
Erkenntnis
https://doi.org/10.1007/s10670-018-0054-0(0123456789().,-volV)(0123456789().,-volV)

1 Denialism as a Social–Epistemic Problem
It is well known that many people feel empowered to reject evidence and expert
testimony on a great number of issues. Such denials range from the absurd and
trivial (that the earth is flat) to the nuanced and dangerous (that the MMR vaccine
causes autism or that the risks of climate change have been overstated). At the
personal level, denial can be ‘‘active’’ or ‘‘passive’’ (Cohen 2001), recognized by the
denier as at odds with the evidence or not. It can be inwardly directed (as in the case
of self-deception or epistemic akrasia) or discursive, as when paid spokespeople
question the science of nicotine addiction (Michaels 2008) or the connection
between HIV and AIDS (Kalichman 2009). Denial can be motivated in various
ways—either by material reward or recognition (Oreskes and Conway 2010a;
Dunlap and McCright 2011) or more subtly by allegiance to one’s ideological tribe,
emotion, or values (Kahan et al. 2011; Markowitz and Shariff 2012; Norgaard
2011).
When active denial has a discursive aim (explicit or implicit), this aim can
likewise vary widely. In the social epistemic context, however, a typical aim is
suspension of others’ belief—or suppression their confidence in that belief—on a
given matter. It has been well documented that the denial of the dangers of cigarette
smoke was intended primarily to forestall regulatory action by seeding doubt on the
relevant science, though the science was well in hand (Michaels 2008; Oreskes and
Conway 2010a). It has thus been common in recent decades to refer to the detractors
of such consensus science as ‘‘denialists’’ or ‘‘contrarians’’ (Oreskes and Conway
2010a; Diethelm and McKee 2009; Kuhn 2012; Michaels 2008; Steffen 2011;
Garvey 2008; Odenbaugh 2012).
Not surprisingly, many of these individuals (and groups) bristle at such monikers,
preferring to be called ‘‘skeptics’’—highlighting, presumably, their putatively
positive epistemic role as counterweights to what they see as an overly dogmatic
scientific community. Science enthusiasts are understandably reluctant to grant
them this label, holding that the sort of ‘‘institutional skepticism’’ we see in science
(Merton 1942) is something importantly different and positive. We see this attitude
expressed in a recent paper by Lewandowski et al.: ‘‘Public debate and skepticism
are essential to a functioning democracy. Indeed, skepticism has been shown to
enable people to differentiate more accurately between truth and falsehood’’ (2016,
538). The ‘‘skepticism’’ promoted by denialists, by contrast, is seen as disingenuous
or ideologically-motivated and thus a distinct phenomenon; Torcello (2016) calls it
‘‘pseudoskepticism’’ (compare the term ‘pseudoscience’).
Lewandowski et al. claim that the ‘‘dividing line between denial and skepticism’’
can be identified ‘‘with relative ease because denial expresses itself with
considerable homogeneity’’ (ibid.), arguing that denialism commonly involves
conspiracy theories, personal attacks, and incubation in ‘‘echo-chambers’’ of various
sorts (538–9) and the avoidance of legitimate ‘‘scientific debate’’ or peer review
(543). But it does not follow from the homogeneity of denialism that it is distinct
from skepticism.
123
M. H. Slater et al.

We do not believe that drawing a sharp distinction between denialism and
skepticism is well motivated. First, the framework that treats skepticism as generally
virtuous and denialism as pernicious is historically and philosophically problematic.
Here we join Coady and Corry in rejecting the view that skepticism is a virtue; as
they put it, ‘‘[s]kepticism about some topics is justified. Skepticism about other
topics is not. Skepticism itself is neither virtuous nor vicious; it should be regarded
as epistemically neutral’’ (2013, 3). Second, we disagree with Coady and Corry in
regarding these denialists, about, say, anthropogenic climate change (henceforth
‘ACC’) as generally aiming for disbelief in some propositions as their discursive
goal. Some, doubtless, are; but as Oreskes and Conway (2010a,b) point out, the
history of organized denial campaigns offers many examples of contrarian
campaigns employing ‘‘contrarian experts’’ to muddy the water and delay action
(e.g., on regulating tobacco or mitigating acid rain). In such cases, while the agents
propounding contrarian claims represent themselves as disbelieving (and may in
fact disbelieve the claims in question), their presumptive aims are satisfied by their
audience merely suspending belief. Lay observers, in a sense, say to themselves:
‘‘Well, some people say P, others say not-P; I don’t know what to believe.’’ As a
strategy for disrupting public consensus formation and thus slowing democratically
supported intervention, this strategy takes advantage of an epistemic asymmetry:
proponents of regulation are playing for the win (belief by sufficient numbers of
citizens) whereas denialists are often playing for a tie (suspension of belief).
One might see this as further reason for thinking that ‘denialist’ paints with too
broad a brush: some alleged ‘‘denialists,’’ after all, do not deny that ACC is
occurring; they merely claim that it’s too soon to tell. Nevertheless, a core unifying
stance for both varieties is the denial that we know (that anyone knows) that ACC is
occuring. As we discuss below, such denial comes by way of discursive strategies
analogous to those of familiar philosophical skeptics. Thus we find it productive to
think of denialism as a certain form of applied skepticism. It is not generally
responsible or epistemically-motivated skepticism, but it is skepticism nonetheless.
It is properly called denial because (skeptical paradoxes aside) scientists do clearly
know that ACC is occurring (IPCC 2014).
Seen in this light, it is surprising that philosophers—particularly those interested
in social epistemology—have devoted little attention to denialism (cf. Torcello
2016). At best, this represents a missed opportunity to engage with a pressing social
issue. If denialism suppresses belief from where it ought to be (given our best
evidence) and if this suppressed belief causes harm, then denialism causes harm. We
submit that both of these ‘‘if’’-clauses are pretty clearly true in general. Concerning
harm, we need only reflect on the efforts of the tobacco industry and the HIV–AIDS
denialism of Thabo Mbecki’s presidency in South Africa to recognize that (almost
certainly) denialism has cost lives. It is poised to continue to do so in the context of
climate change and the question of the safety of childhood vaccines as well (here
too it probably already has).
Psychology has been on the case, accumulating both direct and indirect evidence
that denial campaigns are effective at suppressing belief (see, e.g., Leiserowitz et al.
2013; McCright et al. 2016). Given what is at stake, a pressing question for
researchers in science education and communication (among other fields) is how we
123
Denialism as Applied Skepticism: Philosophical and…

might ‘‘inoculate’’ the laity from the skepticism promoted by motivated denialists
(van der Linden et al. 2017). But so far, little attention has been devoted to
discerning what specific denialist strategies are most effective; in the experiments of
McCright et al, for example, a variety of denialist talking points and challenges
were assembled together to form a ‘‘climate change denial counter-frame’’ that they
found to be effective at suppressing belief in anthropogenic climate change. While
this ‘‘cocktail’’ of challenges was found to be effective, we should not assume that
each specific challenge contributed equally to the suppression of subjects’ belief—
indeed, some may have mitigated the effect of others owing to ‘‘boomerang’’
effects. Assuming that effective ‘‘inoculation’’ and denial-response strategies will
need to be tuned to the specific denialist challenge, it would be helpful to know
which strategies for belief suppression are most effective. One aim of this paper is to
make progress on this goal.
Answering this question empirically requires regimenting denialist strategies in
some way and experimentally testing their efficacy in belief suppression. There are
many ways of doing this. One approach might simply collect a large number of
verbatim talking points commonly found in various forums, conferences, blogs,
books, TV appearances, and so on, and test the most common. We opted for a more
general approach—taking our cue from thinking of denialism as a form of ‘‘applied
skepticism’’—distinguishing strategies on the basis of the generality of their content
and whether they raised questions or issued a specific challenge.
The paper is structured as follows: in the next section (Sect. 2), we distinguish
between a few basic varieties of denialist challenge and compare them to different
traditions of skepticism. In Sect. 3, we describe two empirical studies that explore
the efficacy of different styles of denialist challenge for suppressing belief. Our
results suggest that a disagreement-oriented skeptical model is more effective; Sect.
4offers a general discussion and suggests some next steps for further research.
2 Varieties of Denialist Strategies
Let us now consider some more specific comparisons between denialist strategies
and certain forms of skepticism. We begin with the most familiar: Cartesian
skepticism. The Cartesian skeptic asks an apparently simple question: Can you be
sure that you’re not dreaming right now? Or possibly: Might you not be in the
Matrix? Do you, in general, have any way of ruling out the possibility that you are
massively deceived right now? Such questions, of course, are supposed by the
skeptic to lack good, non-question-begging answers. And if you cannot rule out that
the proffered skeptical scenario is true—even if it seems absurdly unlikely—then
you cannot rightly claim to know anything that such a massive deception would
undermine. Something similar is apparently going on when climate change or AIDS
contrarians pointedly ask whether you can be sure that everything you think you
know about climate change or the HIV–AIDS link isn’t in fact misinformation
spread by a powerful conspiracy. Most of us have little direct, first-person
acquaintance with the relevant evidence; we are almost entirely dependent on the
say so of third parties (Lipton 1998; Keren 2007; Anderson 2011). But if we cannot
123
M. H. Slater et al.

answer this general question about our ability to rule out the possibility of others’
error or dishonesty, then how can we claim to know—and shouldn’t we reduce our
confidence in—the propositions we claim to know?
Notice that in this case, the skeptic need not actually assert that the skeptical
scenario—the conspiracy, massive error, or whatever it is—actually is the case.
They merely need to raise the possibility in a way that makes it salient. To many,
simply opening the question suffices to achieve the skeptical outcome. This can be
observed in a philosophically and scientifically ham-fisted (yet often apparently
effective) way in the context of public debate over matters that are acknowledged to
admit of a certain level of uncertainty. Oreskes and Conway point out in an editorial
in Nature (2010b) that scientists are often overcareful about acknowledging
uncertainties and caveats, ‘‘outlining what they don’t know before proceeding to
what they do—a classic example of what journalists call ‘burying the lead’’’ (687).
Having granted some possibility that a scientific proposition might be wrong, some
are erroneously inclined to think of the situation as an even-odds guess.
Cartesians can press conspiracy theories progressively further, for example by
pointing to purported facts that seem anomalous or in conflict with some ‘‘official
story’’ as suggestive that a conspiracy is afoot. The ‘‘Climategate’’ incident—that is,
the theft and selective quoting of private emails of climate scientists—offered a
particularly rich opportunity for casting doubt on the trustworthiness of climate
scientists in general. By quoting out of context correspondence between scientists
that seemed nefarious to the laity, e.g., concerning the ‘‘manipulation of data’’ (a
phrase that can sound sinister out of context), climate change denialists were able to
raise to salience the possibility of a massive, organized deception without presenting
any evidence that there was such a thing. The question then becomes how can we be
sure that there’s not a massive conspiracy of scientists? Don’t we need to verify that
there is not before we can accept anything they have to say? This can develop into
full-fledged assertions that the conspiracy theory or skeptical scenario in question is
in fact the case. Many well-known denialists of course take this route.
Just as Cartesian skeptics call into question the general reliability of our belief-
forming mechanisms, science skeptics might likewise cast doubt on a scientific
claim by asserting that science is unable to marshall sufficient evidence or be
sufficiently objective to speak reliably on a given topic. One can do this by claiming
that the specific evidence is by its nature unreliable (or cannot be trusted) or by
claiming that the evidence does not exist or is incompatible with other more
trustworthy or probative facts. We might think of these cases as involving very
general challenges—either by raising pointed questions or making stronger
assertions concerning the possibility of general scenarios that would tell against a
large body of claims at once.
We can contrast this broadly Cartesian strategy with an approach adopted by
some Pyrrhonian skeptics. While Pyrrho (via Sextus Empiricus) offered some
general, transcendental arguments for skepticism, one of the Pyrrhonian’s go-to
techniques was to raise specific challenges. Fogelin puts the contrast between the
general and specific strategies nicely:
123
Denialism as Applied Skepticism: Philosophical and…

At first glance, Pyrrhonian skepticism may, indeed, seem mild in comparison
with various forms of Cartesian skepticism. There is something exhilarating,
almost giddy, in the thought that all of our common beliefs about the world
might just be false, and Cartesian skeptical scenarios seem to raise just this
possibility in a vivid form…. [I]t does not take radical—globally dislocating—
scenarios to introduce suspension of belief. It is quite sufficient to note—and
dwell on—the fact that our empirical claims are made in the face of
unchecked, though checkable, defeators…. Given any empirical assertion, it is
always possible—indeed always easy—to point to some uneliminated (though
eliminable) possibility that can defeat this claim. Nothing like brains in vats
are needed to achieve this purpose. (Fogelin 1994, 192–193)
Let us thus consider this aspect of Pyrrhonian skepticism as emblematic of a specific
approach to offering a skeptical challenge (in contrast with the more general
Cartesian approach). As before, we can think of such challenges as proceeding
dialectically either via questions or assertions. For example, leaving the restaurant,
you say that your car is parked on 3rd Street—indeed, you know that it’s parked
there. The skeptic challenges: Are you sure? Do you know, for instance, that it has
not been stolen in the past five minutes? This question works in the same way as the
Cartesian’s question when it comes to this particular knowledge claim; you
probably cannot rule this possibility out and so may fail to feel satisfied that you do
know where your car is parked. But the Pyrrhonian has a dialectical advantage over
the Cartesian here owing to the tradeoff of generality for specificity: the possibility
itself may seem quite a bit more credible. Car theft, after all, is something that
occurs.
A Pyrrhonian employs corresponding assertions when they draw on the skeptical
modes from disagreement (Annas and Barnes 2000, 40–41). If my epistemic peer
offers me an argument pointing to a conclusion at odds with my belief—supposing
the considerations are indeed ‘‘equipoised’’—then it seems that I ought to suspend
belief. This simple idea has been much discussed in recent years under the rubric of
the epistemology of disagreement. In his now classic paper, Christensen offered a
case that many of us can easily relate to:
Suppose that five of us go out to dinner. It’s time to pay the check, so the
question we’re interested in is how much we each owe. We can all see the bill
total clearly, we all agree to give a 20 percent tip, and we further agree to split
the whole cost evenly, not worrying over who asked for imported water, or
skipped dessert, or drank more of the wine. I do the math in my head and
become highly confident that our shares are $43 each. Meanwhile, my friend
does the math in her head and becomes highly confident that our shares are
$45 each. How should I react, upon learning of her belief?
I think that if we set the case up right, the answer is obvious…. I should lower
my confidence that my share is $43 and raise my confidence that it’s $45. In
fact, I think (though this is perhaps less obvious) that I should now accord
these two hypotheses roughly equal credence. (Christensen 2007, 193)
123
M. H. Slater et al.

In Christensen’s case, it’s important that we recognize the disagreeing party as our
epistemic peer. We would probably not feel compelled to suspend our belief (or
even decrease our confidence) if the disagreement was with an eight year old or our
friend who is notoriously bad at math.
Note also the first-person character of the case. Many instances of denialism,
however, will be second-person—for example, when an assertion of one apparent
authority is countered by a denial from another. The shift from a first-person to
second-person context may actually heighten the skeptical impact of this
disagreement. A neutral observer of this disagreement may not be well positioned
to determine who is a relevant epistemic authority and who is not. Doing so can be
difficult enough in non-contentious contexts, but is especially fraught when efforts
are taken to artificially build the apparent credibility of one testifier and reduce that
of the other (Almassi 2012; Collins and Evans 2007; Goldman 2001). Thus a sort of
second-order skepticism can take hold: if I must suspend belief about whether the
disagreeing parties are epistemic peers, then I may not be able to rule out their
disagreement as epistemically significant—for all I know, it is, and I should suspend
my belief.
A second consequence of the disagreement’s being second-person is that the
neutral party often will not be in possession of information that might resolve the
disagreement. In the check-splitting case, the grounds for my answer are
introspectively available to me; in certain circumstances, this may encourage
steadfastness (Kelly 2010; Lackey 2010). Merely observing the conflict, however, I
will generally be deprived of such immediate opportunities to evaluate the grounds
for either position. These effects can also interact: when an apparent expert
challenges a generally acknowledged expert by raising a very specific or technical
point—which neutral observers may be unable to follow—it may also suggest to the
observers that the challenger is a well-informed, epistemic peer of the acknowl-
edged expert.
Worse yet, few of us are actually neutral. When asked to determine whether some
person with a certain range of qualifications is a ‘‘knowledgeable and trustworthy
expert,’’ our answers will often be guided mainly by whether their views fit with our
cultural outlooks (Kahan et al. 2011, 167; Kahan 2014; Ahola 2016). Denialists
adopting this strategy thus rely on motivated cognition while simultaneously
facilitating its operation by offering skeptics of consensus science a plausible
justification for why they ought to suspend their belief.
We could doubtless make other comparisons, but as a start, let us regiment our
denialist strategies as varying along two parameters: form (question vs. assertion)
and content generality (specific or general). These yield four general strategies that
loosely correspond to the discursive skeptical strategies discussed above. One
reason for regimenting strategies this way is the possibility of a fruitful interaction
with anti-skeptical strategies explored in the philosophical context. For example,
one might adopt a sort of Moorean response to the denialist who raises very general
questions about the reliability of science as a method for making contact with the
world (1925/1993, 1939/1993); or one might employ Austin’s ordinary language
approach to the specific question strategy (1946/1979).
123
Denialism as Applied Skepticism: Philosophical and…

Consider the latter for a moment as an example. Suppose that John makes a
specific claim to knowledge—that there’s a goldfinch in the garden, say. You ask
him pointedly ‘‘Are you sure it’s not a goldcrest?’’ Austin points out that there are a
number of specific questions you may have in mind; more importantly, he argues
that there had better be something behind your question if John is to take it
seriously. If John answers that he knows from the markings and you persist by
saying ‘‘Are you sure that another bird doesn’t have those same markings?’’ or
‘‘That’s not enough,’’ Austin thinks that something untoward is potentially going
on: ‘‘you must have in mind some more or less definite lack. ‘To be a goldfinch,
besides having a red head it must also have the characteristic eye-markings’ [and so
on]…If there is no definite lack which you are at least prepared to specify on being
pressed, then it’s silly (outrageous) just to go on saying ‘That’s not enough’’’ (1946/
1979, 84). This point can be neatly connected to Torcello’s (2016) contention that
denialism is ‘‘pseudoskepticism’’ in light of its unwholesome motivations (in
contrast to those adopting a general methodological skepticism familiar in science
or expressing rationally well-founded doubts). The common thread here is that
certain skeptical challenges are legitimately rebuffed or ignored on the basis of their
origin (if, for example, they are disingenuous or better characterized as expressions
of ideology rather than reflections of the evidence).
This is not meant as a defense of Austin’s normative point. But assuming that it is
sound, it suggests an avenue of response to denialists who adopt a certain genre of
skeptical strategies. There would still remain a psychological question of whether
the proffered response would be generally effective in turning away a denialist
challenge. But before researchers address this question, it is worth knowing which
of the four broad classes of skeptical strategies are psychologically more effective at
instilling doubt. As noted above, researchers already have evidence that denialism is
generally effective, but what kind of denialist challenges deserve our attention as the
most effective is so far unclear. Thus we turn to two empirical studies we conducted
to answer this question. These results inform ongoing work determining whether it
is possible to inoculate targets of denialism against the untoward suppression of
their beliefs.
3 The Studies
Here is a more precise statement of our initial research question (discussed in Sect.
3.1): given the communication of a relatively simple scientific idea, which of our
four-fold content–form typology of denialist challenges (general assertion, general
question, specific assertion, specific question) are most effective at suppressing
subjects’ acceptance of the communicated idea? Of course, there are a number of
further dimensions we could have considered (in addition to form and content).
Anderson et al. (2014), for example, studied the effect of ‘‘tone’’ in online
communication fora. One could also examine a more subject-specific regimentation,
identifying certain patterns or talking points that are familiar in the context of a
particular subject (such as climate change or vaccine safety). These may be worth
exploring and the methodology we describe is easily adapted to other approaches.
123
M. H. Slater et al.

Our second research question (discussed in Sect. 3.2) is whether skeptical
challenges exhibit a ‘‘dose–response’’ effect—that is, do more challenges lead to
greater suppressions of belief?
In this set of studies, we chose a subject matter paralleling cases where denialist
efforts have been prominent—in three senses: (1) an issue on which science is
primarily relevant and needed (as opposed to, say, a question of policy); (2) for
which a strong scientific consensus has emerged; and (3) a matter that concerns risk.
However, we also sought to investigate a case that would differ from prominent
targets of denialists in being (so far) relatively non-politically- or ideologically-
entangled and about which many people are largely ignorant. Preliminary testing
indicated that the risks of the Zika virus would fit this profile. This topic is not
apparently ideologically entangled (a finding confirmed by our results), but has
suffered from some misunderstanding in the public—in particular, confusion about
the ways in which the virus is spread, the link between Zika and the birth defect
Microcephaly, and the severity of the disease for those who contract it have been
documented (HORP 2016).
3.1 Four Modes of Denialist Challenge (Study 1)
3.1.1 Methods
3.1.1.1 Participants Study 1 employed a sample of 800 participants from the
American population with a completion rate of 94 percent (final sample size
n = 758). The embedded experiment was conducted online using Qualtrics survey
software and the sample was collected using Amazon Mechanical Turk in February
and March of 2017. Participants were compensated 50 cents for completion of the
survey. All results were analyzed using SPSS.
3.1.1.2 Experimental Design Participants were asked to take part in a study on the
communication of scientific theories. The between subjects experiment began with a
short expository paragraph detailing the leading theory concerning transmission and
risks of the Zika virus (see online Supporting Material). After reading the expository
paragraph, participants were asked to rate their degree of belief in three statements
related to the description of the theory. Belief ratings ranged from zero (completely
disagree) to 100 (completely agree). The belief statements were as follows:
1. The Zika virus can be spread through sexual transmission.
2. The greatest risk from the Zika virus is to pregnant women and their babies.
3. Microcephaly is a condition in which a baby’s head is much smaller than
expected.
Participants were then randomly assigned either a denialist challenge from one of
four challenge categories, or a control statement. Participants in the control group
were told ‘‘The Zika virus theory presented to you was printed in a popular science
magazine’’. The participants from each of the denialist challenge groups read the
statement: ‘‘The Zika virus theory presented to you was printed in a popular science
123
Denialism as Applied Skepticism: Philosophical and…

magazine. The article generated many responses, including the following:’’ In this
way, the challenges were presented as public comments to the Zika theory piece.
The four challenge categories were designed to test the effects of content (general or
specific) and form (question or assertion) (Table 1). All participants were then
retested on the same three belief statements, again selecting a belief rating between
0 and 100.
After completion of the embedded experiment, participants were debriefed and
offered the option to exclude their results from the study. This exclusion, and that of
participants who did not fully complete the study, resulted in a final sample size of
n = 758.
3.1.1.3 Measurement The effectiveness of the denialist challenge statements was
determined by the overall change in total belief scores. Participants scores on the
three belief statements were totaled to create a pre-belief score out of 300, and a
post-belief score out of 300. The difference between these scores (post-belief minus
pre-belief) was the dependent variable of belief change. Therefore, a negative
difference between pre-challenge total belief score and post-challenge total belief
score would constitute an effective denialist challenge. The more negative a belief
change score, the more a participant’s belief lowered between measurements.
Demographic variables included were Age (M = 37, SD = 12.3, Range =
18–75), gender (50.3% female), level of education (49.5% with bachelor’s degree
or higher), income (M = $50,000), race (83.2% white), and political party
affiliation (42% democrat). The results of all demographic measurements can be
seen in the Supporting Material. In addition to these demographics, participants
were asked to respond to two value predispositions. Religiosity was measured by
asking ‘‘How important is religion in your life?’’ Responses were recorded on a 5
point scale with 1 = ‘‘Not at all important’’ and 5 = ‘‘Extremely important’’
(M = 2.36, SD = 1.49). Trust in science was measured by asking ‘‘How much
would you say you trust in the findings of science?’’ with 1 = ‘‘Not at all’’ and
5 = ‘‘A great deal’’ (M = 4.11, SD = .855). We included religiosity and political
ideology to confirm that we had, in fact, chosen a subject (Zika) that was not
significantly tied to religious or political ideologies. Trust in science and religiosity
are frequently negatively correlated (Nadelson and Hardy 2015, Cacciatore et al.
2016), and we were also interested in the ways these variables might interact with
Table 1 Denialist challenges used in Study 1
Form
Content Question Assertion
General Do scientists really have enough evidence to
support these theories?
Scientists don’t have enough evidence to
support these theories!
Specific Can’t you only get Zika through a mosquito
bite?
You can only get Zika through a mosquito
bite!
123
M. H. Slater et al.

belief change. We predicted that those with high trust in science would have high
pre-belief scores, and would be less likely to lower their belief.
3.1.2 Results and Analysis
A one-way between subjects ANOVA for belief change (post-belief minus pre-
belief) showed a significant difference between the five treatment groups
F(4753) = 10.26, p\.001. Mean belief change for each of the treatment groups
can be seen in Fig. 1, below. Post hoc analysis using Tukey’s HSD
1
indicated that
the ‘‘Specific Assertion’’ condition (M = -10.20, SD = 25.03) and the ‘‘General
Assertion’’ condition (M = -13.44, SD = 40.58) both produced significantly more
change in belief than the control (M = 3.91, SD = 21.99). Neither the ‘‘Specific
Question’’ condition (M = -1.69, SD = 23.84) nor the ‘‘General Question’’
condition (M = -1.54, SD = 19.16) were significantly different from the control.
Those in the ‘‘Specific Assertion’’ condition were also shown to have a significantly
greater change of belief than those in the ‘‘General Question’’ condition, and the
‘‘General Assertion’’ condition participants had a significantly greater change in
belief than those in the both the ‘‘General Question’’ and ‘‘Specific Question’’
conditions. In Fig. 1you can also see that those in the control condition showed an
increase in belief. This is expected as an exposure-effect, as subjects saw the same
group of belief questions twice.
To more explicitly test the effects of form versus content, we conducted a two-
way ANOVA for the effects of content (‘‘general’’ vs. ‘‘specific’’ denialist prompts)
and form (‘‘question’’ vs. ‘‘assertion’’ denialist prompts) on belief change. There
was no statistically significant interaction between the effects of content and form
on belief change, F (1753) = .586, p= .444, and no significant difference in belief
change for the content groups (p= .484). There was, however, a significant
difference in belief change between form groups (p\.001).
A one-way between subjects ANOVA comparing belief change between all those
participants receiving a ‘‘general’’ denialist prompt, all participants receiving a
‘‘specific’’ denialist prompt, and the control group revealed a significant difference
between groups F(2755) = 9.32, p\.001. Post hoc analysis with Tukey’s HSD (see
note 1) showed that both ‘‘specific’’ (M = -5.94, SD = 24.77) and ‘‘general’’
(M = -7.53, SD = 32.30) prompt groups changed their belief significantly
compared to the control group. Both types of content were, therefore, similarly
effective.
In contrast, post hoc analysis of the significant one-way between subjects
ANOVA comparing belief change between the ‘‘question’’, ‘‘assertion’’, and control
groups (F(2755) = 20.03, p\.001) revealed that only the ‘‘assertion’’ group
(M = -11.83, SD = 33.75) was significantly different from the control, while the
‘‘question’’ group (M = -1.62, SD = 21.59) was not.
For all participants receiving denialist prompts, no significant correlations were
found between belief change and religiosity, and no significant differences were
1
Tukey’s HSD was chosen as the appropriate post hoc test because the sample sizes are close to equal
with similar variances and Tukey has good power and control over Type 1 error.
123
Denialism as Applied Skepticism: Philosophical and…

found for political party affiliation. A significant relationship was found between
trust in science and pre-belief, r = .279, and belief change, r = .159, (both
ps\.001), indicating that participants with higher trust in science had a higher pre-
belief, and were less likely to change their beliefs in the face of denialist challenges.
3.1.3 Discussion
The correlation between trust in science, pre-belief, and belief change indicate that,
as expected, the more trust one has in science, the more inclined one is (though
obviously not perfectly so) to accept messages presented as scientific conclusions
and to retain these beliefs in the face of denialist challenge. The results of Study 1
indicate that denialist prompts worded as assertions rather than as questions are
more effective at lowering belief in a non-ideologically entangled scientific claim.
Although the general and specific assertions targeted different aspects of the given
theory, they were still more effective at lowering belief than the specific and general
questions. Statistical analysis of the difference in mean belief change between
treatment groups, and those examining the differences in forms versus the effects of
content support lead to a consistent finding. At least in the case of the Zika virus,
form has a much more important influence on the effectiveness of denialist claims
than the specificity (or lack thereof) of the content. This suggests that whether
denialists formulate their skeptical challenges in grand, Cartesian generalities or
Pyrrhonian specifics, they are well-advised to do so in the stronger, assertive forms.
Skeptical efficacy thus seems to cut across skeptical traditions (at least as we
rudimentarily regimented them).
Fig. 1 Mean belief change for treatment groups in Study 1; * = significance at p\.001 level
123
M. H. Slater et al.

3.2 Dose–Response and Interaction Effects (Study 2)
Our second study built on the results of Study 1 by focusing on the assertion
challenges (we dropped the question challenges). Our aims in this study were to
(a) replicate our initial observation that counter-assertion was effective in
suppressing belief in large demographically-representative sample, and (b) deter-
mine whether this suppression exhibited a dose–response effect.
3.2.1 Methods
3.2.1.1 Participants Study 2 employed a nationally-representative sample of 1802
participants from the American population (collected by Research Now in March of
2017) with a completion rate of 83 percent. The embedded experiment was
conducted online using Qualtrics survey software and participants were compen-
sated $2.50 in e-rewards currency.
3.2.1.2 Experimental Design As in Study 1, participants were asked to take part in
a study on the communication of scientific theories. The between subjects
experiment began with the same short expository paragraph detailing the leading
theories of the Zika virus (see the online Supporting Material). After reading the
expository paragraph, participants were asked to rate their degree of belief in the
same three statements used in Study 1, with ratings again ranging from zero
(completely disagree) to 100 (completely agree). Study 2 participants were then
randomly assigned either one denialist challenge, three denialist challenges, or a
control statement.
Given the results of Study 1, three specific assertions and three general assertions
were used as challenges in Study 2 (Table 2). The specific assertions included the
one used in Study 1, as well as two designed to target common misconceptions
about Zika (HORP 2016). The general statements included the same general
assertion used in Study 1, and two general statements that could be applied to any
scientific theory about which members of the public might perceive risk.
Participants receiving one challenge (Dose 1 group) were randomly assigned one
of the six challenges. Participants receiving three challenges (Dose 3 group) were
assigned a random combination of 3. All possible combinations of challenges
yielded 20 ‘‘treatment sets’’ (as we shall call them) and one control group (all Dose
3 treatment sets can be seen in Fig. 3). The challenges were again presented as
public comments to the Zika theory piece, which the all participants were told had
appeared in a popular science magazine. All participants were then retested on the
same three belief statements, again selecting a belief rating between 0 and 100.
3.3 Measurement
The effectiveness of the denialist challenge statements was determined by the
overall change in total belief scores in the same way as was calculated in Study 1.
Participants scores on the three belief statements were totaled to create a pre-belief
123
Denialism as Applied Skepticism: Philosophical and…

score out of 300, and a post-belief score out of 300. The difference between these
scores (post-belief minus pre-belief) was the dependent variable of belief change.
Demographic variables included were Age (M = 50.64, SD = 14.55, Range =
19–94), gender (57.3% female), level of education (51.6% with bachelor’s degree
or higher), income (M = $60,000), race (87.4% white), and political party affiliation
(34.2% democrat). The results of all demographic measurements can be seen in the
Supporting Material. The two value predispositions, Religiosity (M = 3.03,
SD = 1.45) and trust in science (M = 4.02, SD = .865), were measured with the
same variables as Study 1.
3.3.1 Results and Analysis
A one-way between subjects ANOVA for belief change showed a significant
difference between the three Dose groups (i.e., between Control, Dose 1, and Does
3): F(2,1799) = 8.18, p\.001. Post hoc analysis with Hochberg’s GT2 revealed
that those participants in the Dose 3 group (M = -11.44 SD = 43.48) lowered their
belief significantly more than those in the Dose 1 group (M = -6.39 SD = 43.48)
and those in the control group (M = .3152 SD = 31.49) (Fig. 2).
Table 2 Denialist challenges for Study 2
Specific S1. You can only get Zika through a mosquito bite!
S2. Zika symptoms aren’t mild at all; they’re life-threatening!
S3. There’s no link between Microcephaly and Zika!
General G1. There isn’t enough evidence to support this theory!
G2. Alarmists are exaggerating this so-called ‘‘problem’’!
G3. There’s no scientific consensus about this theory!
Fig. 2 Mean belief change across dose groups in Study 2. * = significance at the p\.001 level
123
M. H. Slater et al.

Beyond the three Dose groups, we sought to identify which particular denialist
statements, or sets of denialist statements, were most effective at lowering belief. A
one-way between subjects ANOVA for belief change showed a significant
difference between the 20 Sets and one control. Post hoc analyses with Hochberg’s
GT2
2
showed that only one of the Dose 3 treatment sets (S3,G1,G3) significantly
lowered belief compared to the control (See Fig. 3and Table 3, found in the
Supporting Information). Given the difference in sample sizes between the Dose 1
Sets and the Dose 3 Sets, separate ANOVAs (both significant) for belief change
were run for all Dose 1 sets together (F(6,996) = 5.49, p\.001) and all Dose 3 sets
together (F(20,962) = 2.31, p\.001). Post hoc analyses were performed for each
ANOVA using Gabriel’s PCT, as the sample sizes within each Dose treatment were
only slightly unequal. Table 3 in the supplemental material shows the means and
standard deviations of all of the sets. Figure 3below shows the means for all sets,
with significance level (as compared to the control) is indicated with asterisks. In
these analyses, Statement 1 from the Dose 1 Set significantly lowered belief
compared to the control.
For those participants who received denialist challenges, belief change was
slightly related to religiosity, r=-.049, p\.05 (this negative correlation indicates
that as religiosity goes up, people are more likely to lower belief) and more strongly
related to trust in science, r =.091, p\.001 (as trust in science increases, people are
less likely to lower belief). Trust in science is also significantly related to pre-belief,
r= .324, p\.001. Again no significant differences were found for political party
affiliation.
3.3.2 Discussion
The initial results of Study 2 comparing the Dose 1 condition and the Dose 3
condition support the idea of a ‘‘dose–effect,’’ in which increasing the number of
denialist challenges to which a participant is exposed causes them to further lower
their belief. This is not, in itself, very surprising. However, it is notable that not
every trio of challenges resulted in a suppression of belief. Indeed, only one of the
three-dose sets led to a statistically-significant drop compared to Control. The non-
significant changes may be explained, in part, by the fact that some of the sets
contained statements that were in some tension with one another: two of the general
statements, G2 and G3, could be seen as opposed to S2. As shown in Fig. 3(observe
dose sets marked with ‘#’), these sets barely lowered or even raised belief (and the
effect was not statistically significant; see Supporting Material for precise values
and standard deviations). One potential exception to this trend is the set S1,S2,G3
(though the effect was still not statistically significant), but the overall pattern seems
to be that consistency matters in reducing layperson confidence; we return to this
matter in the conclusion.
2
The design of Study 2 included multiple levels of sorting that rendered some sample sizes unequal.
Hockberg’s GT2 was the appropriate post hoc analysis in this case given the differences in sample sizes
between the control group and the Dose 1 and Dose 3 sets.
123
Denialism as Applied Skepticism: Philosophical and…

One specific statement in Study 2 was more effective compared to the others in
the single-dose group (S1: ‘‘You can only get Zika through a mosquito bite!’’). One
possible explanation for S1’s relative effectiveness is that it plays into the most
common misconception held by the public about Zika (HORP 2016). Yet (internal
conflicts aside), S1 does not seem to drive any obvious pattern in the multiple-dose
sets. As mentioned above, one combination of statements (S3,G1,G3) was
statistically-significantly effective at suppressing belief. It remains unclear what
exactly about this combination subjects found so compelling.
In this larger sample size, participants were slightly more likely to lower belief if
they indicated that religion was more important to them. This relationship was small
but significant. As in Study 1, as participants’ trust in science increased, they were
less likely to lower belief in response to denialist claims.
4 Conclusion and Next Steps
To summarize, we’ve found that—at least for the present case—denialist strategies
were notably more effective when they worked in the mode of disagreement as
opposed to questioning the grounds of subjects’ beliefs. One can imagine several
explanations for this observation. Here is one: Perhaps simply raising a question
about one’s facts (or about the reliability of one’s sources or belief-forming
methods) without explanation fails to motivate the possibility that one’s facts might
really be wrong (or unreliable), absent specific reason for thinking that there’s a
problem. This resonates with Austin’s general anti-skeptical approach: you might
question my facts or credentials, but unless you have in mind ‘‘some specific lack’’
Fig. 3 Mean belief change detail (Study 2); * = significance at the p\.001 level; # = conflicting dose
sets
123
M. H. Slater et al.

such questioning often seems to fall flat. Compare two versions of Christensen’s
dinner check-splitting example: in the original version, when your friend arrives at a
different conclusion than you do it seems quite natural to suspend (or lower) your
belief that you are correct; but if your friend merely asked you ‘‘Are you sure that
it’s that much?’’ it seems both normatively and psychologically plausible that you
should (or may) remain steadfast. You might reply ‘‘Yes, I just did the math. Do you
have some reason for thinking I’m wrong?’’ Suppose now that your friend replies
‘‘Well, yes: I just did the math too and got an answer $5 lower.’’ This arguably
changes the dynamic (though it might not always change our response). When your
friend asserts a different answer, they essentially provide the sought-after grounds
for their question.
Whereas Study 1 suggests that denialist challenges that had the form of assertions
were more effective in suppressing belief than the question equivalents—more than
any differences in the generality of a challenge’s content—Study 2 shows that (at
least in the present context) content matters. While we do see a dose–response effect
for one combination of challenges (and overall), this was not true for each trio of
challenges. This is consistent with the results of McCright et al. (2016) in which a
number of distinct denials at various levels of specificity were woven together into
an all-purpose ACC ‘‘counter-frame’’. But simply increasing the number of
challenges did not reliably decrease the belief suppression.
Not all patterns in these data are readily explained, but the hypothesis floated
above—that perceived tensions within the three-dose sets of challenges reduced
their efficacy—goes some ways to explain why more denial didn’t generally lead to
more skepticism. Again, we can imagine versions of our check-splitting scenario
that illustrate the psychological plausibility of both of these effects. Four people set
out to calculate the correct share of the dinner check. Your confidence wavers when
you learn that one of your friends has arrived at a different number; it might drop
significantly (if you’re like us) when you learn that all three check-computers agree
on the correct value—especially if they’ve arrived at it independently. Alterna-
tively, imagine that while your friends disagree with you about the correct value,
they also disagree with each other. Now, perhaps, it’s not as clear what the proper
response is. It seems plausible to us that the drop in confidence should not be quite
as pronounced. Remaining steadfast in one’s own calculation seems psychologically
more tenable in this case than it is in the previous.
In the context of denialist challenges to consensus science, we can imagine a
number of different ways in which tensions/inconsistencies might manifest:
(synchronic-individual) on the same occasion by the same person; (diachronic-
individual) on different occasions by the same person; (synchronic-group) on the
same occasion (roughly) by different people; (diachronic-group) on different
occasions by different people. Presumably cases like (synchronic-individual) will
tend to have an undermining effect on the undermining power of the such
challenges. It’s not as obvious that (diachronic-individual) would—or should. The
details should matter. People change their minds—responsibly in some cases. Does
this look like such a case or is it more plausibly a forced retreat to a different
position consistent with a predetermined position they will defend come what may
(as in ACC-denialists who shift from denying warming to denying the human-
123
Denialism as Applied Skepticism: Philosophical and…

causation claim to denying that ACC is bad at all)? Similar complexities apply in
the last two cases as well. It is an open question what considerations influence the
efficacy of denialist strategies and why. Is it typical for members of the lay public to
employ the sort of inference-to-the-best-explanation approach Lipton (1998)
suggests we often use in evaluating testimony generally? To what extent do
cognitive biases mitigate (or accentuate) the otherwise compromising effect of
inconsistency?
Returning to the big picture, our studies suggest that research into strategies for
forestalling denialism’s effects ought to focus in the first instance on the more
‘‘disagreement-oriented’’ analysis of denialism. Are there ways in which members
of the public can be ‘‘inoculated’’ from this form of denial in initial communication
and education efforts? Are there ways in which the skeptical effects of counter-
messaging can be reversed after the fact—e.g., by getting subjects to recognize
counter-messaging that is motivated by ideology or material gain (Christensen
2014), perhaps by pointing out suspicious inconsistencies or non-epistemic
explanations of the disagreement (e.g., generous funding by interested parties or
particular policy preferences)? There are clearly rich opportunities for philosoph-
ical–psychological collaboration on such projects. Philosophical accounts of the
normative appropriateness of steadfastness might be used to inform further
psychological investigation to determine whether we may apply such norms in
real world contexts.
It is also an open question at this point whether our results generalize to other
cases. As discussed, we intentionally chose a case that showed little interaction with
political or religious ideology in order to focus on the epistemic phenomena. This
may provide insight into prospective use of denialist skeptical strategies but
interactions with ideology will presumably complicate matters in cases about which
there is already significant denialist noise. But how exactly this works and how it
can be combated are important questions that deserve philosophers’ and psychol-
ogists’ attention.
References
Ahola, S. (2016). Why (not) disagree? Human values and the readiness to question experts’ views. Public
Understanding of Science, 26(3), 339–354.
Almassi, B. (2012). Climate change, epistemic trust, and expert trustworthiness. Ethics & the
Environment, 17(2), 29–49.
Anderson, E. (2011). Democracy, public policy, and lay assessments of scientific testimony. Episteme,
8(2), 144–164.
Anderson, A. A., Brossard, D., Scheufele, D. A., Xenos, M. A., & Ladwig, P. (2014). The ‘‘nasty effect:’’
online incivility and risk perceptions of emerging technologies. Journal of Computer-Mediated
Communication, 19(3), 373–387.
Annas, J., & Barnes J. (Eds. and Trans.) (2000). Sextus Empiricus: Outlines of scepticism. 2nd Ed.
Cambridge: Cambridge University Press.
Austin, J. L. (1946/1979) Other minds, reprinted in his Philosophical papers. Oxford: Oxford University
Press.
123
M. H. Slater et al.

Christensen, D. (2007). Epistemology of disagreement: The good news. Philosophical Review, 116(2),
187–217.
Christensen, D. (2014). Disagreement and public controversy. In J. Lackey (Ed.), Essays in collective
epistemology. New York: Oxford University Press.
Coady, D., & Corry, R. (2013). The climate change debate: An epistemic and ethical enquiry.
Basingstoke: Palgrave Macmillan.
Cohen, S. (2001). States of denial: Knowing about atrocities and suffering. Cambridge: Polity Publishers.
Collins, H., & Evans, R. (2007). Rethinking expertise. Chicago: University of Chicago Press.
Diethelm, P., & McKee, M. (2009). Denialism: What is it and how should scientists respond? European
Journal of Public Health, 19(1), 2–4.
Dunlap, R. E., & McCright, A. M. (2011). Organized climate change denial. In J. S. Dryzek, R.
B. Norgaard, & D. Schlosberg (Eds.), The Oxford handbook of climate change and society. Oxford:
Oxford University Press.
Fogelin, R. J. (1994). Pyrrhonian reflections on knowledge and justification. New York: Oxford
University Press.
Garvey, J. (2008). The ethics of climate change: Right and wrong in a warming world. London:
Continuum Publishers.
Goldman, A. I. (2001). Experts: Which ones should you trust? Philosophy and Phenomenological
Research LXIII, 1, 85–110.
HORP (2016) Public views of the Zika virus Outbreak topline data. In Harvard T. H. Chan School of
Public Health: Harvard Opinion Research Program.
IPCC. (2014). Climate change 2014: Synthesis report. Contribution of working groups I, II and III to the
fifth assessment report of the intergovernmental panel on climate change. Geneva: IPCC.
Kahan, D. M. (2014). Making climate-science communication evidence-based—all the way down. In D.
A. Crow & M. T. Boykoff (Eds.), Culture, politics and climate change: How information shapes our
common future (pp. 203–220). Abingdon: Routledge.
Kahan, D. M., Jenkins-Smith, H., & Braman, D. (2011). Cultural cognition of scientific consensus.
Journal of Risk Research, 14(2), 147–174.
Kalichman, S. C. (2009). Denying AIDS: Conspiracy theories, pseudoscience, and human tragedy. New
York: Copernicus Books.
Kelly, T. (2010). Peer disagreement and higher-order evidence. In R. Feldman & T. A. Warfield (Eds.),
Disagreement (pp. 111–174). Oxford: Oxford University Press.
Keren, A. (2007). Epistemic authority, testimony and the transmission of knowledge. Episteme, 4(3),
368–381.
Kuhn, J. (2012). Passive smoking and ‘‘denialism’’: Motives and strategies of denial of scientific
evidence. Pra¨vention, 35, 11–14.
Lackey, J. (2010). A justificationist view of disagreement’s epistemic significance. In A. Haddock, A.
Millar, & D. Pritchard (Eds.), Social epistemology. Oxford: Oxford University Press.
Leiserowitz, A. A., Maibach, E. W., Roser-Renouf, C., Smith, N., & Dawson, E. (2013). Climategate,
public opinion, and the loss of trust. American Behavioral Scientist, 57(6), 818–837.
Lewandowsky, S., Mann, M. E., Brown, Nicholas J. L., & Friedman, H. (2016). Science and the public:
Debate, denial, and skepticism. Journal of Social and Political Psychology, 4(2), 537–553.
Lipton, P. (1998). The epistemology of testimony. Studies in the History and Philosophy of Science,
29(1), 1–31.
Markowitz, E. M., & Shariff, A. F. (2012). Climate change and moral judgement. Nature Climate
Change, 2, 243–247.
McCright, A. M., Charters, M., Dentzman, K., & Dietz, T. (2016). Examining the Effectiveness of
climate change frames in the face of a climate change denial counter-frame. Topics in Cognitive
Science, 8, 76–97.
Merton, R. K. (1942). The normative structure of science, reprinted in his the sociology of science:
Theoretical and empirical investigations. Chicago: University of Chicago Press.
Michaels, D. (2008). Doubt is their product: How industry’s assault on science threatens your health.
Oxford: Oxford University Press.
Moore, G. E. (1925/1993) Defence of common sense. In T. Baldwin (Ed.), G. E. moore: Selected
writings. London: Routledge.
Moore, G. E. (1939/1993) Proof of an external world. In T. Baldwin (Ed.), G. E. Moore: Selected
writings. London: Routledge.
123
Denialism as Applied Skepticism: Philosophical and…

Norgaard, K. M. (2011). Living in denial: Climate change, emotions, and everyday life. Cambridge, MA:
The MIT Press.
Odenbaugh, J. (2012). Climate, consensus, and contrarians. In W. P. Kabasenche, M. O’Rourke, & M.
H. Slater (Eds.), The environment: Philosophy, science, and ethics. Cambridge: MIT Press.
Oreskes, N., & Conway, E. M. (2010a). Merchants of doubt. New York: Bloomsbury Press.
Oreskes, N., & Conway, E. M. (2010b). Defeating the merchants of doubt. Nature, 465, 686–687.
Steffen, W. (2011). A truly complex and diabolical policy problem. In J. S. Dryzek, R. B. Norgaard, & D.
Schlosberg (Eds.), The Oxford handbook of climate change and society. Oxford: Oxford University
Press.
Torcello, L. (2016). The ethics of belief, cognition, and climate change pseudoskepticism: Implications
for public discourse. Topics in Cognitive Science, 8, 19–48.
van der Linden, S. L., Leiserowitz, A., Rosenthal, S., & Maibach, E. (2017). Inoculating the public
against misinformation about climate change. Global Challenges,1(2), 1600008.
123
M. H. Slater et al.