Society

https://doi.org/10.1007/s12115-022-00670-7

BOOK REVIEW

Steven Pinker, Rationality: What it Is, Why it Seems Scarce,

Jonathan Dancy 1

# The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2022

This book, which is effectively a written version of a class one thinks that a patient is 90% likely to have cancer, things
Pinker has been giving at Harvard for several years now, has that are invasive, expensive and potentially dangerous.
many strengths. It is extremely well written, its main charm So Pinker is full of missionary zeal in this respect. The
being that you really feel you are getting to know the author as more people actually understand how probability works, the
you go along. There is an engaging person behind the text. better off we will all be.
And classes always (or almost always!) go better if the instruc- But the book is called Rationality, and this, as I see it, is
tor is having a good time, and Pinker is plainly enjoying him- where the problems begin. Rationality, as Pinker understands
self here. He sweeps you along with his enthusiasm and ener- it, is the capacity to reason; and to reason is to move from
gy. But there is also an evangelistic element in the mix. Pinker considerations adduced to a conclusion of one sort or another,
thinks that if everyone knew, and more importantly under- that is, to infer. There are two species of reasoning: reasoning
stood, the things he is telling us, then everyone would be better to belief and reasoning to action. Call these Theoretical
off, the world would be safer and we could all worry a lot less. Reasoning and Practical Reasoning. Practical Rationality,
Of course much of what he is telling us is not really new. we are told, is the capacity to work out how you are most
Our appalling and apparently ineradicable tendency to make likely to get what you want most, which you do by reasoning.
mistakes about probabilities, to which a lot of space is devot- Pinker doesn’t have a great deal to say about this type of
ed, is well known – to the experts, if not to practitioners like reasoning, basically, I suppose, because he thinks it is not very
you and me. But it doesn’t matter whether it is new or not. It is interesting. It is just ordinary reasoning for a practical purpose,
important anyway. One of his examples: if people knew more in the service of desire. Theoretical Rationality, which oc-
about the probabilities involved, they (or at least those who cupies much more of the book, is the capacity to follow, or
can afford it) would invest more in their pensions, just for a to construct, inferences. There are two sorts of inference: for-
start. mal and probabilistic. Formal reasoning guarantees the truth
Here is a well known and particularly striking instance of of its conclusion if the premises are true; probabilistic reason-
our tendency to make such mistakes, one among many (from ing establishes the degree of probability of the conclusion
p. 150). Suppose that the prevalence of breast cancer in the given the truth of the premises. Pinker runs us through a basic
population of women is 1%. And suppose that the sensitivity course on formal reasoning, but this isn’t his real concern,
of a breast cancer test (its true-positive rate) is 90%. Suppose since people don’t tend to make mistakes about that. It is
that its false-positive rate is 9%. A woman tests positive. What probabilistic reasoning that is the real bugbear since it is so
is the chance that she has the disease? Pretty much everyone hard to get it right. Accordingly, we have one chapter on
says that the answer is 90%. But the correct answer is 9%. And formal reasoning and six on probabilistic reasoning. The book
this is not just of academic interest: a lot of things get done if ends with two chapters about why people hold the crazy be-
liefs they do and why it is so hard to get them to (see reason
and) change their mind.
Pinker scampers through the chapter on formal reasoning
* Jonathan Dancy without looking right or left. The well-known oddities of the
[email protected] propositional calculus are given no stress. For instance, ‘if p
1 then q’ is understood in the propositional calculus as ‘not both
University of Texas at Austin, Austin, TX, USA
 Soc

p and not-q’. On this understanding of ‘if’, it is true that if New find out what one already wants. Suppose I have accepted a
York is in the USA, Prince Philip died recently. But most of us job as Dean of my Faculty. Starting in, I ask myself what my
would think that these things have nothing to do with each aims should be in this new post, with all the opportunities and
other, so that neither follows from the other. What this shows pitfalls it presents. What should I be most concerned about and
is that formal logic has no way of capturing the notion of one what should I not be concerned about or not so much at least? I
thing following from another – the notion of a consequence. decide that my primary concern should be equity, which is a
But it manages to get on perfectly well (in its own terms) decision about what to care about most. And this is a decision
without that. about what to want.
Now this is not really a criticism of formal logic; it is what So far, then, Pinker has dealt with formal reasoning and
it is and not another thing. Theoretical Rationality is among probabilistic reasoning, both of which are reasoning from be-
other things the capacity to follow arguments expressed in lief to belief, and with practical reasoning, where we are rea-
formal terms. It is also the capacity to follow arguments soning from belief to action, seeking to make choices consis-
expressed in probabilistic terms. Reasoning of this sort, prob- tent with our beliefs and with our other choices. But does
abilistic reasoning, is complex and difficult. Many of us are rationality consist merely in the capacity to construct or follow
very bad at it, and I am no exception. Pinker gives an excellent arguments – to reason, in that sense?
account of this, as far as I am capable of judging though I One way into this large and difficult question is to ask what
don’t think there is anything being said here that is new. is distinctive of rational creatures like us. What is it that we can
I want to return to the distinction between practical and do that makes us non-trivially distinctive? We reason, yes, but
theoretical rationality. Suppose that theoretical rationality is there is something even more basic than that: we are capable
the capacity to move from beliefs to beliefs in certain reputa- of recognizing reasons in ways that have nothing to do with
ble ways. So understood, I reason theoretically when, starting reasoning understood as inference. Our ability to recognise
from beliefs that my daughter said she would be home by reasons and to respond to reasons as such is surely as much
midnight and that the last train has now arrived without her, a central and distinctive element of our rationality as is our
to the belief that she is taking a taxi. One might have thought ability to infer. And there is absolutely nothing in this book
that practical rationality is, analogously, the capacity to move about our capacity to recognise and respond to a reason. I
from beliefs to action (or at least to an intention to act, if not to regard this as a notably large shortcoming in a book about
the action intended). But according to Pinker, practical ratio- rationality. It also means that when we come to the last two
nality is not that: it is the capacity to determine how one is chapters, where Pinker is trying to come to terms with the
most likely to get what one most wants. But this looks like a recent explosion in ideologically motivated craziness (in cer-
move not to action but to belief – to a belief about probabili- tain parts of the world), we are deprived of one essential tool,
ties, in fact. And if so, action can never itself be the product of the idea that people take a consideration that is not a reason at
reasoning in the way that belief can be. Practical reasoning is all to be a reason of some strength, and take considerations
at most reasoning to a conclusion that has practical relevance that are at best only very weak reasons for belief as clear
(to us!). But I would want to know why we should allow in indications or even as proof.
advance that an action can never stand in the sorts of relation What is a reason, then? It is admittedly hard to say anything
to considerations adduced in reasoning that belief can stand in. that is not circular or banal about this. I find myself saying that
Why cannot an action be undertaken in the light of complex a reason is a consideration that counts in favour of a response.
considerations which together make a case for acting in that Responses can be practical – if the reasons are reasons for
way? My daughter has not come home with her brother as action – or theoretical –if they are reasons for belief. We are
usual, and the only way she can now get home is by train; so I (probably) the only creatures we know of that act for reasons –
drive to the station. The ‘so’ here seems to be the same ‘so’ as that respond to considerations as favouring our so responding,
the ‘so’ in ‘so she is taking a taxi’. i.e. believe what we see reason to believe and do what we see
In addition to this, there might also be what one might call reason to do. We alone can respond to reasons as reasons. And
desiderative reasoning, which is reasoning aimed at determin- don’t forget that deciding that the time is ripe for running is as
ing an aim (or desire). Wants and aims do not just assail us; we much an action as deciding to run, and as running too. Of
can form them, and do so for reasons. So at this point my course many animals do what they have good reason to do.
thought is that something significant has been missed out in But it appears they are not capable of acting in the light of
this instrumental conception of practical reasoning. It is as if those considerations as reasons so to act. This is what is prob-
wants are just given. You find yourself with them, and the ably distinctive about us.
only question thereafter is how to manoeuvre to obtain max- So though rationality includes the capacity to act for a
imal desire satisfaction. The question what to want is one that reason, there is nothing whatever about that aspect of it in
can hardly be addressed. But there is such a thing as trying to Pinker’s Rationality. I view this as a significant criticism be-
work out what to want. And this is not the same as trying to cause there is a strong suggestion that the book covers the
Soc

whole topic rather than just one aspect of it. As evidence of the voted for Donald Trump, causing great consternation all
intellectual blinkers being worn here, I offer the remark in round. But nobody supposed that this impugned their grasp
Chapter 1 about the African Bushmen, or San, who sometimes on the canons of rationality. Their logical capacities are ex-
help their enemies or rivals when they are in deep distress. ceptional.) So they are reasoning, yes, but their reasoning will
Pinker says that they do this because memories are long and be what Pinker calls ‘motivated reasoning’. This is a strange
fortunes may reverse, so that the San’s reason for helping is a term because reasoning is an activity and as such likely to be
self-interested calculation. But this is not the only possible engaged in for a reason (such as the desire to find the truth) –
explanation. Another is the moral explanation: they help be- but what Pinker means by it is reasoning motivated by some-
cause, though these people are rivals, they are in dire straits thing other than the simple desire to find the truth. So people
and need help. I would want to know why this is not a possible reason in ways that lead to conclusions that suit them. This has
explanation – and note that, if it is, it is nothing to do with always been the case, not very surprisingly; what is surprising
calculating how best to get what one most wants (the only sort is the way in which it is our ideological (especially political)
of practical reasoning that Pinker allows). And, as I have al- preferences that have come to determine so much of what we
ready suggested, that one wants something is not just a brute think.
fact. There is always the question what to want to be born in Pinker’s move here is to distinguish between what he
mind. Desires do not (always) just arrive and impose calls the reality mindset and the mythology mindset. The
themselves on us. We can decide what is worth wanting and reality mindset is concerned with how things really are in
pursue it in that light. And in doing so we are dealing with our immediate surroundings, the people we deal with face
reasons. to face, our factual memory of previous interactions, and
This is where moral reasoning comes in. Suppose that I the rules and norms that regulate our lives (p. 200). This is
have a debt which comes up for payment next week, but the rather a strange list, because rules seem so different from
person who lent me the money has absolutely no need for it facts. But Pinker thinks that most people get such matters
yet. If I don’t repay her next week, I can use the money to help right and reason perfectly rationally about them. Then
someone get home to see his mother before she dies. I can’t there is the world beyond ‘immediate experience’, the dis-
get in touch with my creditor – if I could things would be easy. tant past and the unknowable future, how things might
My question is: what do I have most moral reason to do? (or have been, the microscopic and the cosmic. Pinker says
just: what should I really do?) Here I have to weigh up the that people have no way of finding out what is or is not
relative importance of paying what I owe according to my the case in these zones, and the question whether this or
agreement, and helping someone see his mother for the last that belief about them is true or false is the wrong question.
time. This is an instance of a sort of moral reasoning that does The function of such beliefs is not to guide us to the truth,
not appear in Pinker’s book. And it is not a form of inference. but to construct a social reality that binds our tribe or sect
The last two chapters are an attempt to understand the ap- (p.300). This is the mythology mindset, and it operates
parently recent phenomenon of ideologically driven craziness. according to different rules. This is the realm within which
The first of these is entitled ‘What’s wrong with people?’. As people believe crazy things.
Pinker recognizes, this is where interest peaks: in large num- To me it seems that Pinker has given up far too much right
bers, people (other people, of course, not rational people like at the start. First, beliefs about the future are as factual as
us!) believe crazy things. (So we are still thinking about the- beliefs about the immediate past. And it is not the case that
oretical rationality here.) Apparently three quarters of we can never know what caused what. We should not give up
Americans believe in something paranormal such as posses- our own claims to knowledge about such matters in order to
sion by the devil (42%), extra-sensory perception (41%), explain the extraordinary beliefs that other people have, and
ghosts and spirits (32%), astrology (25%), witches (21%), we should not be forced to allow that because some people
communicating with the devil (29%) and so on.1 All this quite have crazy views, the question whether it is they or we who
apart from the peculiarities of recent political events in the US are right, or even whether there is something for us to be right
(and not only in the US, alas), birtherism, anti-vax – once you about, is the wrong question.
start the list seems endless, and it has got longer recently. So Moreover, even if beliefs about what might have happened,
what is going on? or would have happened in different circumstances, are in
Pinker insists that we must allow that people who have some good sense not factual (they are after all commonly
beliefs we think are crazy still have the capacity to grasp called counterfactual) this does nothing whatever to show that
‘the canons of rationality’ (so rationality must have canons the function of such beliefs is to construct a social reality that
for us to grasp). (I have two colleagues in my Faculty who binds our tribe or sect. The disjunction Pinker postulates is too
extreme. The function of beliefs about the safety and efficacy
1
https://news.gallup.com.poll/16915/three-four-americans-believe- of the vaccine is not to bind our tribe. We have been given too
paranormal.aspx few options.
 Soc

Now Pinker recognizes that as children of the Enlightenment My final point is that Pinker’s focus on inference may be
we hold that all our beliefs should fall within the reality mindset. partly the cause of his difficulties here. We should return to
And submitting all our beliefs to the test of reason and evidence reasons. Ideologically motivated beliefs are not held for the
is an unnatural and at times painful skill that has to be learnt. reason that they subserve one’s ideology. There is a ‘because’
Schools and universities can help by teaching courses on statis- here which is not the ‘because’ of ‘for the reason that’. They
tical and critical thinking. Still, he says, there is a side of our are held for ordinary reasons, but (in many cases) inadequate
nature that resists it. We resist the demand that all our beliefs ones. An ideological motivation may nevertheless explain
should be held to this standard. We can have commitments that why one takes what is not much of a reason for a very strong
are not beliefs; otherwise put, we can have commitments that we reason, or fails to see the strength of a very strong reason – for
don’t think are factually true but which are very important to us. example, the reason(s) to be vaccinated.
What is most obvious, and most worrying, is that ideological
opinions are increasingly determining where one stands on mat-
ters that should be independent of ideology. A classic example
Publisher’s Note Springer Nature remains neutral with regard to jurisdic-
is the reading of statistics in the attempt to grasp what sort of tional claims in published maps and institutional affiliations.
gun control would be most effective; people make mistakes that
align closely with (and seem to be best explained by) their
political opinions. The same applies to views about vaccination Reviewer: Jonathan Dancy, FBA is Professor of Philosophy at the
and many other topics. The hard question here seems to be how University of Texas at Austin and author of several books, including
Practical Reality (2000), Ethics Without Principles (2005) and
ideological distinctions have been allowed to become so central Practical Shape: A Theory of Practical Reasoning (2018).
in all this. I don’t think that we have yet got much of an answer
to that challenge or how to counteract it.