by Hugo Mercier, Dan Sperber
‘The Case Against Reality’ was among my favourite books last year. As almost an aside somewhere, it goes, ‘we did not evolve our ability to reason in order to pursue the truth. We evolved it as a tool of social persuasion.‘ Quite a statement to make, and I was intrigued. The Enigma of Reason expands on this, and with the rigour of scientific reasoning (no pun intended) has me quite convinced.
I, like most others, have considered our species’ power of reasoning to be the superpower that got us here. And yet, when I think about, it seems paradoxical because our decisions also tend to have a lot of biases, and are flawed. In The Enigma of Reason, Hugo Mercier and Dan Sperber seek to resolve this.
In doing that, they challenge one of our deepest assumptions about human thinking: that reason evolved to help individuals make better decisions and discover truth. Instead, they argue that reason’s real purpose is social. It evolved not to make us smarter in isolation, but to make us more persuasive and cooperative within groups.
Through five parts, the authors propose an interactionist approach to reason as against the (above) intellectual approach. That is not to say that it does not bring intellectual benefits, but that’s almost a side effect. The main story is social.
The Enigma of Reason begins by questioning the “intellectualist” view that portrays reason as a general-purpose, truth-seeking faculty, and the more recent dual process (System One, Two) version. Instead, it postulates that reason is a specialised inference module – not a general engine for all forms of thought, but a mechanism that draws intuitive inferences about reasons themselves.
It operates much like the visual system: highly specialised, yet capable of indirectly influencing a wide range of domains. Just as vision processes light patterns to infer the nature of objects, reason processes arguments to infer justifications.
This specialisation allows for what they call “virtual domain-generality,” meaning that although reason is narrowly tuned, it indirectly contributes to thinking about almost any topic. In the case of reason, all the inferences we make involve modules, and ‘representations’ – material things that give give us information about the actual/desired state of affairs. Inferential procedures erase/modify/create new representations from existing ones.
The book also quickly brings up the paradox I mentioned earlier – we are often biased, irrational, and stubborn even when evidence contradicts our belief/thought. We cherry-pick data that fits our views, rationalise poor decisions, and fall prey to confirmation bias. If reason was meant to help us find truth, it seems badly designed. The book flips this puzzle superbly, like that old product joke – what if it’s a feature, not a bug?
That leads to the central theme of the book, called the interactionist theory of reason, which suggests that reasoning developed to serve communication and social coordination. Early humans lived in cooperative groups where survival depended on collaboration and trust. In such contexts, the ability to justify one’s actions, convince others, evaluate arguments offered by peers, and being able to set expectations of self and others (by giving reasons, we commit to a future line of thought and conduct), would have been far more valuable than detached introspection.
Reason, therefore, evolved as a tool for argumentation – to produce reasons that defend our views (justification) and to assess those of others (evaluation). It follows that the main role of reason is not to guide/motivate us in reaching conclusions, but to explain and justify after the fact the conclusions we have reached.
From this perspective, confirmation bias stops looking like a defect. At a ‘production/ justification’ level, when individuals search for reasons supporting their beliefs, they are biased and lazy, but discussions get them to generate arguments that can be tested and criticised by others. From a ‘consumption/evaluation’ perspective, perspective, we tend to be more objective and demanding.
The Enigma of Reason calls this process “epistemic vigilance” – our instinct to scrutinise others’ claims for signs of deception or error. Group discussion thus acts as a filter that weeds out weak reasoning. The social exchange, rather than solitary reflection, is where truth can emerge.
This is backed with research from psychology, anthropology, and evolutionary biology. The book describes experiments showing that individuals reason poorly when alone but perform much better when debating in groups – provided there’s open dialogue and accountability. The implication is striking: humans reason best when they reason together.
The Enigma of Reason also explores how this view explains modern phenomena like political polarisation and social media debates by highlighting when reason is taken out of its social context, it potentially misfires. In echo chambers, where dissent is minimal, the argumentative function of reason breaks down. Without genuine opposition, people simply reinforce their existing beliefs – a malfunction of the mechanism that evolved to thrive on challenge and exchange.
In the end, it also provided explanations for a niggling doubt I had – the individual brilliance of scientists. Through examples of individuals (Heisenberg, Newton), and groups (The Royal Society) the book shows how science historically has used the best features of reasoning – discussions, peer reviews, and tools/methods (labs, replications, forums) to facilitate well-crafted arguments as well as change their minds. Even the greatest scientists have made mistakes. Lastly, unlike many other professions, scientists – at their best – strive for truth and not the approval of an audience.
By redefining reason as a social adaptation as opposed to simply an intellectual tool, The Enigma of Reason does a radical refresh on (at least my) understanding of human cognition – that rationality is less about lone enlightenment and more of collective improvement – that progress, science, and even morality emerge from our argumentative nature, which filters our individual biases.
In terms of accessibility, The Enigma of Reason goes into the weeds right at the beginning after the introduction. I did find it difficult – good stand alone thoughts but not giving me a forest view. If you’re willing to go through about a 100 pages of that, this book is rewarding and deeply thought-provoking, with an excellent ‘Conclusion’ section that wraps it all up very well. In my Bibliofiles 2025 list.
Notes & Quotes from The Enigma of Reason
1. I have thought of reasoning as this one meta mechanism/logic which is a sort of Swiss knife across different kinds of contexts and types of problems. But that’s what animals have. In humans it goes beyond the ‘instinctual’ and into acquired territory. There are many specialised mechanisms to draw inferences and reason is only one of them. In fact there is a continuum running from inference to intuition to reason. That ‘mechanism’ we perceive happens because reason draws inferences/conclusions from the reasons we come up with for various contexts and in doing so, indirectly produces reflective conclusions about the things reasons are themselves about – anything from physical objects to living beings to concepts like numbers, law etc.
2. Why did reason only develop in humans? Therein lies the first clue – humans occupying a very specific ecological niche only humans inhabit. Requiring co-operation with not just kin, but strangers too, for short and long terms. (I think this is great input for the origin of language as well, beyond simpler animal versions of communication)
3. Cognition is a means for organisms that move to react appropriately to risks and opportunities presented by the environment. Goes beyond what is understood by senses, and gets into inferences
4. Instinct to expertise is also a continuum. These mechanisms are modular. May have distinct evolutionary and development trajectories. Individual modules are rigid but articulation of different modules provides complex organisms with adaptive flexibility. Evolved and expert skills exploit specific brain areas, and when these are injured, skills are impaired
5. All Inference involves representations (fascinating that this is also the first step in Reshuffle, in the context of machine learning)
6. The reason why some original reasons are weak is because of the trade-off between effort and payoff.
7. We sometimes make sub optimal decisions, because they are easier to justify to others!
8. Jonathan Haidt wrote “The Emotional Dog and its Rational Tail” – reasoning is here only to create post-hoc justifications that cover the tracks of the intentions and emotions secretly running the show. He even suggested that instead of making us do the right thing, reason may gives us excuses not to do the right thing! And it’s proven in experiments. 😐 Related to why people do not change moral opinions based on rational reasons. Haidt says “like thinking that forcing a dog’s tail to wag by moving it with your hand will make the dog happy”
Leave a Reply