by Hugo Mercier
slave through a demonstration of the Pythagorean theorem.
Socrates doesn’t have to force any conclusion on the slave: once
each premise is presented in a proper context, the slave can draw
the appropriate conclusions himself. Socrates only has to or ga-
nize the steps so that the slave can climb them on his own. In a
way, the answer has been “given out [by the slave] of his own
head,”13 and yet the boy would likely never have reached it with-
out help.
This illustrates the efficiency of reasoning as a mechanism of
open vigilance. Reasoning is vigilant because it prompts us to
54 ch ap t er 4
accept challenging conclusions only when the arguments reso-
nate with our preexisting inferential mechanisms. Like plausibil-
ity checking, reasoning is essential y foolproof. Typical y, you
receive arguments when someone wants to convince you of
something you wouldn’t have accepted other wise.14 If you’re too
distracted to pay attention to the arguments, you simply won’t
change your mind. If, even though you’re paying attention, you
don’t understand the arguments, you won’t change your mind
either. It’s only if you understand the arguments, evaluating
them in the pro cess, that you might be convinced.
Reasoning not only makes us vigilant but also open- minded,
as it helps us accept conclusions we would never have believed
without argument. I mentioned previously studies showing that
people tend to put more weight on their own views than on other
people’s opinion, moving only about a third of the way toward
the other on average (the length of the Nile example). When
people are provided with a chance to discuss the issue together,
to exchange arguments in support of their views, they become
much better at discriminating between the opinions they should
reject and those they should accept— including opinions they
would never have accepted without arguments.15
The vast majority of our intuitions are sound— other wise we
wouldn’t be able to navigate our environment, and we would
have been selected out a long time ago. Because we recruit these
intuitions to evaluate the reasons people offer us, the reasons we
recognize as good enough to warrant changing our minds should,
more often than not, lead to more accurate opinions and better
decisions. The exchange of arguments in small discussion groups
should thus tend to improve per for mance in a variety of tasks,
as people figure out when to change their minds, and which new
ideas to adopt. This is exactly what has been observed, as ex-
changing reasons al ows forecasters to make better predictions,
w h at t o b e l ie v e ? 55
doctors to make better diagnoses, jurists to make better judicial
decisions, scientists to develop better hypotheses, pupils to bet-
ter understand what they’re taught, and so on.16
Challenging Arguments
To make us more open- minded, reasoning should evaluate ar-
guments as objectively as pos si ble. In par tic u lar, people should
be able to spot a good argument even if its conclusion is deemed
implausible. When we’re told that the correct answer to the
Linda, Paul, and John prob lem is Yes, plausibility checking says
no (for people who got the prob lem wrong). When your col-
league suggests taking the bus, which you know to be slower,
plausibility checking says no. Yet in both cases, once the argu-
ments have been presented, reasoning does its job, overcoming
the initial rejection suggested by plausibility checking. But maybe
you would have been less likely to accept the arguments if you
had felt more strongly that their conclusion was wrong?
In an experiment with participants solving the Linda, Paul,
and John prob lem, we got participants so confident in the wrong
answer that we had to add some extra options on the confidence
scale; other wise, they all claimed to be “very confident” they had
chosen the right answer. Even with the extra options, we ended
up with many participants saying they were “as confident as in
the things I’m most confident about” that the wrong answer was
correct. Yet these ultraconfident participants were just as likely
to recognize the right argument, when it was provided to them,
as those who were less confident.17
Arguments might change people’s minds when they relate to
riddles. For all their overconfidence, people aren’t particularly
attached to their (wrong) answers. But what about things that
matter: our personal life, politics, religion? Are we still able to
56 ch ap t er 4
evaluate arguments objectively then? Three strands of evidence—
experimental, historical, and introspective— make me hopeful
that good arguments generally change people’s minds, even when
they challenge deeply held beliefs.
In many experiments, participants have been offered argu-
ments that vary in strength— from the downright fallacious to
the unimpeachable— and asked to evaluate them. In other stud-
ies, researchers have mea sured how much the participants
changed their minds as a function of the quality of the arguments
they were presented with. The conclusion of these studies is that
most participants react rationally to arguments, rejecting falla-
cious arguments outright, being more convinced by strong than
by weak arguments, and changing their minds accordingly.18
Historical evidence also suggests that arguments work, even
when they support conclusions that are quite revolutionary. In
the early twentieth century, some of the West’s greatest minds—
such as Bertrand Russel , Alfred North Whitehead, and David
Hilbert— were attempting to provide a logical foundation for
mathe matics. In 1930, a young, unknown mathematician named
Kurt Gödel offered a proof that this could not be done (more
precisely, that it is impossible to find a complete and consistent
set of axioms for all of mathe matics).19 As soon as they read the
proof, every one who mattered accepted it, even if that meant
jettisoning de cades of work and kissing their dream good- bye.20
Beyond mathe matics and its perfect demonstrations, good ar-
guments also work in science, even if they challenge established
theories. It is simply not true that, as Max Planck complained,
“a new scientific truth does not triumph by convincing its
opponents and making them see the light, but rather because
its opponents eventually die, and a new generation grows up
that is familiar with it.”21 When the evidence is there, new
theories are promptly accepted by the scientific community,
w h at t o b e l ie v e ? 57
however revolutionary they may be. For example, once suffi-
cient evidence had been gathered to support plate tectonics, it
only took a few years to turn it from fringe theory to textbook
material.22
Good arguments even work in the po liti cal and moral realm.
In The Enigma of Reason, Dan Sperber and I reviewed the exam-
<
br /> ples from mathe matics and science just mentioned but also the
amazing story of British abolitionism, when a nation was per-
suaded to abandon the slave trade in spite of the economic costs
involved.23 In many countries, the past de cades have seen dra-
matic improvements in the rights of women, LGBT+ individu-
als, and racial minorities. In each of these cases, community
leaders, intellectuals, journalists, academics, and politicians spent
time and effort developing arguments drawing on a variety of
moral and evidential sources. People read and heard these argu-
ments and appropriated some of them in their everyday con-
versations.24 Even if good arguments are not the only cause of
the momentous changes we have witnessed, the efforts spent
developing, laying out, and communicating good reasons for
people to change their minds have likely contributed to dramatic
changes in public opinion.
On a personal level, I believe we have all experienced, at some
time or other, the pull of uncomfortable arguments. When I was
going to university, being a (staunch) leftist was very much taken
for granted. But I kept encountering arguments that challenged
some of the po liti cal tenets widely accepted among my peers.
Ignoring these arguments would have had no negative practical
consequences for me personally— then as now, I have essentially
no po liti cal power— and would only have brought social ben-
efits in the form of approval by my peers. Yet I couldn’t help but
feel the strength of these alternative arguments. Even if their con-
clusions remained for some time somewhat disturbing to me,
58 ch ap t er 4
they played a significant role in shaping my current po liti cal
opinions.
Arguably, it is this power of a wel - thought- out argument that
led Martin Luther to develop a distaste—to put it mildly— for
reason, distaste that he expressed quite floridly: “Reason is by
nature a harmful whore. But she shall not harm me, if only I re-
sist her. Ah, but she is so comely and glittering. . . . See to it that
you hold reason in check and do not follow her beautiful cogita-
tions.”25 In the context of the religious battles he was fighting,
we can imagine Luther encountering arguments that challenged
his moral and religious views. Had Luther been able to easily re-
ject these arguments, had he found no strength in them at al ,
he surely would not have felt this internal turmoil and developed
such resentment toward reason.
What If Our Intuitions Are Wrong?
When it comes to evaluating the content of communicated in-
formation, people rely on two main mechanisms: plausibility
checking, which compares the content of the message with our
preexisting beliefs, and reasoning, which checks whether argu-
ments offered in support of the message resonate with our pre-
existing inferential mechanisms.
I have made the case that plausibility checking and reasoning
function well as devices of open vigilance. We can rely on our
prior beliefs to evaluate what we’re told without falling for a con-
firmation bias or becoming more polarized because of backfire
effects. Reasoning evolved to allow a greater degree of open-
mindedness, as good arguments help people accept conclusions
they would never have accepted other wise, even conclusions
that challenge deeply held beliefs.
w h at t o b e l ie v e ? 59
The main issue with plausibility checking and reasoning isn’t
their potential biases in relating communicated information to
our prior beliefs and inferences but the prior beliefs and infer-
ences themselves. Our minds evolved so that most of our beliefs
would be accurate, and most of our inferences sound. Indeed, in
the wide variety of domains we evolved to deal with— from fig-
uring out what food to eat to what people mean when they
speak—we perform wel . We also draw sound inferences in do-
mains that are evolutionarily novel, if we’ve had a lot of oppor-
tunities for learning: bil ions of people are now able to read with
near- perfect fluency; human computers— before they were re-
placed by electronic versions— could perform complex mental
calculations with few slipups.
By contrast, if we attempt to draw inferences in any domain
that evolution and learning haven’t equipped us to deal with, we
are likely to be systematically wrong. When tackling novel ques-
tions, we grope for solutions and reach for some adjacent cogni-
tive mechanism that appears relevant to the question at hand.
This adjacent cognitive mechanism might well be the same for
most of those struggling with the same question. When many
people get things wrong in the same way, cultural patterns can
emerge.
Imagine someone with no access to science wondering why
animals have traits that are so well adapted to their environment.
We’re not equipped with mechanisms to answer this question
directly (why would we? the practical benefits are essentially nil).
By contrast, recognizing and understanding artifacts matters
a lot, and we’re likely equipped with cognitive mechanisms to
deal with that prob lem. Artifacts, in their own way, are also
adapted to their environment. Because we know that artifacts
are made by agents, it appears to make sense that adaptive traits
60 ch ap t er 4
in animals have also been created by agents.26 This is one of
the reasons why creationism has a field day, as people find it
intuitively compelling— certainly more compelling than the
Darwinian idea that adaptations appear through an unguided
pro cess of natu ral se lection.
The same logic applies to many popu lar misconceptions.
Take vaccination. In the popu lar imagination, it consists in tak-
ing healthy babies and injecting them with something that con-
tains a bit of disease (in fact, the agents are typically inert). All
our intuitions about pathogens and contagion scream folly.27
The recent rise in anti- vaccination sentiment has often been
blamed on specific people, master persuaders— from Andrew
Wakefield in the United Kingdom to Jenny McCarthy in the
United States— who would have swayed significant swaths of
the population into harmful, scientifically il iterate be hav ior. In
fact, the anti- vaccination movement is as old as vaccination.
Already in 1853, England’s enactment of the first Compulsory
Vaccination Act “provoked enormous fears of contamination.”28
After a lull in the early twentieth century— likely due, at least
in part, to the plainly vis ible success of the polio vaccine— fears
about vaccination have rebounded in the West. These fears
create a demand for anti- vaccination rhe toric, demand that is
rapidly met. As historian of medicine Elena Conis argued,
“Both the Wakefield study [linking, falsely, vaccines and au-
tism] and McCarthy’s prominence as a vaccine skeptic [are] the
products— not the cause—of today’s parent
al vaccine wor-
ries.”29 The Wakefield study, in par tic u lar, only had an impact
on the U.S. rate of MMR vaccination (the vaccine that had been
fraudulently linked to autism) around the year 2000, when it
was known only by professionals, before it was found to be fake
and its results contradicted by dozens of other studies. By con-
trast, the media frenzy that surrounded the vaccines–autism
w h at t o b e l ie v e ? 61
link, starting a few years later, had no effect on vaccination
rates.30
The examples can be multiplied. We’re not well equipped to
think about the politics or economics of large, complex, diverse
states. Instead, we have recourse to intuitions that evolved by
dealing with small co ali tions in conflict with each other.31 These
intuitions tell us, for instance, that if someone else benefits too
much from trading with us, then we must be losing out, or that
we should be wary of power ful enemy co ali tions colluding
against us. Hence the popu lar success of protectionist or, more
generally, antitrade policies, and of conspiracy theories (obvi-
ously, these intuitions are sometimes right: some trade is bad
for us, and some people do conspire).
Not all misconceptions can be neatly explained by the idea
of misapplied intuitions. The French are big on homeopathy, and
I’m still stumped as to why— how is duck liver diluted up to the
point that there’s nothing left of it supposed to cure the flu? In
chapter 8 I will mention some decidedly nonorthodox beliefs
shared by the inhabitants of Montaillou in thirteenth- century
France— that preserved umbilical cords could help win lawsuits,
for example. They also baffle me (seriously, if you have sugges-
tions, let me know).
Stil , compared with, say, creationism, re sis tance to vaccina-
tion, or conspiracy theories, these odd misconceptions are much
more culturally specific (the umbilical cord thing in par tic u lar
hasn’t caught on wel ). Some explanation for their existence is
certainly required, but the point stands that popu lar miscon-
ceptions, as a rule, are intuitively compel ing. In the absence of
strong countervailing forces, it doesn’t take much persuasion to
turn someone into a creationist anti- vaxxer conspiracy theorist.
Far from being due to widespread credulity, the prevalence of
intuitive misconceptions reflects the operation of plausibility