Reasoning is funny. Not the act of reasoning-although it can also be fun-but the cognitive skill that allows us to figure out math problems and decide what computer to buy. We can agree that reasoning is responsible, at least in part, for some of the greatest achievements of humankind, from calculus to the International Space Station. Yet we should also agree that people can reason their way to asinine beliefs -- we are poisoned by the souls of aliens dropped in volcanoes millions years ago -- and disastrous decisions -- Napoleon's ‘invasion' of Russia. And it's not that some people are smart while others are dumb: Isaac Newton spent more time bent on alchemy than on mathematics and, for their many faults, neither L. Ron Hubbard nor Napoleon were dunces.
Psychologists interested in reasoning cannot expect to recreate such dramatic events as the creation of calculus or the decision to attack Russia in the lab. Instead, they must work with ‘toy models,' simple problems that people can try to solve in less than half an hour before going back to their routine. The most commonly used of these problems is known as the Wason Selection Task -- after its creator the pioneering psychologist Peter Wason. Many of you may already be familiar with it, but if you aren't, you can try to solve the problem yourself - it's laid out in figure one. Take your time.
Did you answer A, or A and 4? If yes, you are in good company, comfortably sitting with the large majority of the thousands of people who've faced the Wason Selection Task. But psychologists would not be nearly as interested in a problem that most people get right: the correct answer, arrived at by about 10% of participants, is A and 7. Why 7? Because if there's a vowel and the other side of the 7 card, then the rule is false. Why not 4? Because even if there's a consonant on the other side of the 4 card, the rule can still be true: it doesn't say that a consonant must have an odd number on the other side. Here's a nice toy model of reasoning's failures: a simple task leading to abysmal performance.
Now for the toy model of reasoning's successes. It is called the Wason Selection Task -- after its creator the pioneering psychologist Peter Wason. It is depicted in figure one. A large majority of people manages to find the correct solution, showing how efficient reasoning can be at solving logical problems.
Wait. How can the very same problem illustrate both reasoning's failures and its successes? The task is the same. The people solving it are the same. What's the catch? The context. In the first toy model, people face the problem on their own. In the second they solve it in groups.
The Wason Selection Task may be a toy model but the boost in reasoning performance offered by group discussion seems to be a fairly robust phenomenon -- one that will be explored at length in future posts (for the impatient, here's a good chunk of the evidence).
We can then ask why reasoning leads to better results in a social than in an individual context. Such a broad question can receive more than one answer. To see that more clearly, it may be easier to ask a similar question about an organ. For instance, why do human lungs lead to better results in an atmospheric than in an aquatic context? A physiologist could look at the structure of the lungs and notice the lack of mechanisms allowing them to extract oxygen from water - hence their failure in that context. The physiologist provides us with a proximal answer: she tries to figure out how the lungs work in these different environments.
Yet even when the physiologist has done her job, one may still be left to wonder why human lungs are not equipped with a mechanism to extract oxygen from water. After all, it's possible: fish are fairly good at it. Clearly the answer to that question is that human lungs evolved in a terrestrial, not aquatic, environment. Their function is to extract oxygen from air, not from water. And so we can expect them to be adapted to the former task, not the later. What we have now is an ultimate explanation: an explanation of why a biological mechanism has been endowed by evolution with certain features.
Back to our puzzle about reasoning: why does it lead to better performance in the context of group discussion? To provide a proximal explanation, we would have to look inside people's head: what happens there that make them better at solving problems when they discuss with others than on their own? In future posts, I will suggest such an explanation. But for now let's focus on the ultimate question. Whatever the psychological mechanisms are that make people better reasoners in group discussion, why do they work that way?
To answer the ultimate question, we can turn to a suggestion made by Dan Sperber in a couple of papers from the early noughties. His idea, in a nutshell, is that reasoning evolved for argumentation: so that we can convince others and to examine the arguments they offer us. Reasoning would be adapted to work in dialogue, when people exchange arguments, and not within the confines of a solitary mind. Just as human lungs work better in normal atmospheric conditions because they are designed to work in these conditions, reasoning works better in group discussion because it is designed to work in such a context.
The performance boost in group discussion is only one of the many puzzles raised by reasoning. One of the main purposes of this blog is to explain how thinking of reasoning as being designed for a social task can help solve these puzzles. But it will also go beyond that. Reasoning is not the only cognitive mechanism to have a ‘hidden' social function. In fact, I will argue that quite a few baffling human behaviors are better explain if we see the mind as designed in large part to deal with the social world. As we will also see in later posts, the big idea is not new, but many twists to it are.
The Wason Selection Task was first presented in this chapter:
Wason, P. C. (1966). Reasoning. In B. M. Foss (Ed.), New Horizons in Psychology: I (pp. 106-137). Harmandsworth, England: Penguin.
The good results in groups were first demonstrated by Moshman and Geil:
Moshman, D., & Geil, M. (1998). Collaborative reasoning: Evidence for collective rationality. Thinking and Reasoning, 4(3), 231-248.
The classical distinction between levels of analysis (four in total) is there:
Tinbergen, N. (1963). On aims and methods in ethology. Zeitschrift für Tierpsychologie, 20, 410-433.
The idea that reasoning evolved to serve an argumentative function was first suggested by Dan Sperber in these papers:
Sperber, D. (2000). Metarepresentations in an evolutionary perspective. In D. Sperber (Ed.), Metarepresentations: A Multidisciplinary Perspective (pp. 117-137). Oxford: Oxford University Press.
Sperber, D. (2001). An evolutionary perspective on testimony and argumentation. Philosophical Topics, 29, 401-413.
Since then, he and I have worked on the theory, joined by other colleagues like Hélène Landemore. Here are the most relevant articles for today's discussion:
Mercier, H., & Landemore, H. (in press). Reasoning is for arguing: Understanding the successes and failures of deliberation. Political Psychology.
Mercier, H., & Sperber, D. (2011). Why do humans reason? Arguments for an argumentative theory. Behavioral and Brain Sciences, 34(2), 57-74.
Warning: unfortunately, posting will have to be relatively infrequent.
This post is the first part of a "reasoning" series.