Menu Close
We can learn to use our minds better by becoming more familiar with how they work. JohnGreenaway

Good reasoning needn’t make you an unfeeling robot

Some interesting recent research using neuroimaging gives us evidence that different brain systems activate in different reasoning situations.

But before we get to that, try the following puzzles:

Puzzle 1: Two playing cards are face up on the table, and two cards are face down. Of the two face-down cards, one has a blue back and the other, red. Of the face-up cards, one is an Ace and the other a King. Which cards must you flip over in order to check that all red-backed cards are Aces?

Puzzle 2: Four people are at the bar. They’re all drinking. You can see that one is drinking mineral water, and another is drinking beer, but you can’t tell their ages. Of the other two, one is clearly under 18 and the other is clearly middle-aged, but you can’t tell what they’re drinking. Who do you need to check to make sure that everyone is complying with the rule that if you drinking alcohol, you’re over 18?

Studies have repeatedly shown people are far more likely to get puzzles such such as number two correct (around 75%) than puzzles such as number one (between 4% and 25%). What explains the difference?

It’s striking that we are good at applying our reasoning in one case but not so good in the other. As someone who studies logic using the tools of philosophy and mathematics, I can see the puzzles have the same logical structure.

They form the basis of the famous Wason Selection Task – a popular test of logical reasoning in experimental psychology.

A user’s guide to networking

Now back to the recent research, published at the end of last month in the journal NeuroImage.

There are two brain networks, called in the literature the “Default Mode Network” and the “Task Positive Network” – and it was shown these activate in different reasoning situations, but rarely together.

One network lit up when subjects were asked to reason about physical systems (including the mechanical properties of inanimate objects); the other lit up when subjects were asked to reason about social situations (including the mental states of other people).

The researchers – from Case Western Reserve University and Washington University in the US – postulate that the two brain systems inhibit each other, which makes it difficult to reason as effectively in tasks encompassing both domains.

Some people have jumped to bad conclusions on the basis of this evidence, claiming that it shows “analytic thinking” and “empathy” are in tension, and that when we reason carefully, we can’t see the human cost of our decisions.

The study doesn’t show anything of the sort. Empathic thinking involves reasoning and analysis as much as any other kind of thinking. After all, there is a reason that psychoanalysis involves analysis.


The NeuroImage data shows we use different neural systems when we process different kinds of information in our thought. It doesn’t show tension between thinking and feeling, or between reasoning and moral judgement.

Notice that in the Wason Selection Task it is the social, personal situation where people seem to reason better, not the abstract impersonal one. We are reasoning in both contexts.

Experiments such as these are important. They can tell us useful things about our capacities, in much the same way other medical research into brains and bodies reveals important things about what we can and can’t do.

But this new paper shouldn’t make us worry that learning to reason is going to make us callous unfeeling robots, any more than getting physically fit might make us thoughtless brutes. Instead, it should encourage us to learn to use our minds better by becoming more familiar with how they work.

Think about it this way …

To think well, we need to do two different but related things.

1) We can all improve our skills at doing reasoning by:

  • learning what good reasoning looks like: both by paying attention to good reasoning when you see it, but also by beginning to spot patterns of reasoning that come up again and again

  • spotting common errors, such as equivocation (when terms you use change in meaning from one part of your reasoning to another) or everyday logical gaffes such as affirming the consequent

  • being able to shift perspectives and look at things from different points of view. If you’re trying to find evidence for some conclusion, stop for a moment and think of what would show that that conclusion was wrong? How would things look then?

  • by growing our store of examples and analogies. Being able to spot how the case you’re looking at might be similar to one you’ve seen before will help you see possibilities you wouldn’t have considered otherwise.

Good reasoning isn’t just a matter of linear steps following programmed rules. It’s also coming up with possibilities to consider and options to explore.

Simply put: the more tools we have at our disposal, and the more comfortable we are with them, the better we’ll reason. If you are used to thinking in different ways, you’ll be able to choose the thinking tool that suits. This is what we do when we teach people to think critically and creatively.

2) We need to better understand what our ways of thinking can do, and what they’re good for. What are the general limits of logical tools? What is the best way of reasoning about morality, or about decision-making and many other things besides?

These are all open questions where the logician, the linguist and the philosopher enter the picture, to help us understand how we can represent and reason about the world. We are the researchers who are curious about people’s conceptual capacities, about what we can do with different forms of representing and reasoning, and where their limits lie.

When we join forces with our colleagues in laboratories who figure out what happens in our brains as we think and feel, we will all get a better picture of how we can engage with each other and the world around us.

How did you go in the two puzzles at the start of this article? If you answered “red card” and “the King” for Puzzle 1, and “the under 18 year old” and “the beer drinker” for Puzzle 2, well done!

Want to write?

Write an article and join a growing community of more than 170,900 academics and researchers from 4,739 institutions.

Register now