Wason Selection Task and the Social Psychology of Reasoning: From Letters and Numbers to Bar Rules

What is the episode about

The Rest Is Science crew explores the Wasson selection task, a cornerstone in the psychology of reasoning, and how framing the problem in abstract (letters and numbers) versus social (bar patrons) terms changes people’s performance. They walk through the logic behind the task, discuss modus tollens and modus ponens, and connect these ideas to broader questions about human rationality, counterexamples, and the social purpose of reasoning.

Key takeaways

• The original Wasson task highlights a counterintuitive flaw in human reasoning when faced with abstract rules.
• Framing a problem around social norms or duties (deontic reasoning) makes counterexamples easier to spot than in purely descriptive logic.
• Reasoning may have evolved as a social tool, not as a truth-seeking mechanism, a view championed by Mercier and Sperber.
• Counterexamples are more informative than confirmations for testing rules, a theme echoed in philosophy of science and epistemology.

Introduction to Wasson’s selection task

The episode centers on Wasson’s selection task, a classic reasoning puzzle introduced in 1966 by Peter Cathcart Wasson. The test presents four cards, each with a letter on one side and a number on the other. The rule to test is: if there is an A on one side, there is a 7 on the other. The hosts illustrate the puzzle, discuss common intuitive mistakes, and contrast the original abstract version with a beer-and-age bar version, showing how framing changes our responses.

"The if part is called the antecedent. It comes before the end, which is called the consequent." - The Rest Is Science

Bar version versus abstract formulation

In the bar version, the cards encode ages and drinks, and the rule becomes: anyone who drinks must be over the drinking age. The bar framing elicits a much more intuitive understanding for many people, who can quickly identify which cases to check (the 12-year-old and the beer drinker) while almost ignoring others. The discussion emphasizes how difficult the abstract version is compared to the social version, even among highly educated participants.

"We evolved to focus on duties and obligations that we invented ourselves and not on universal, timeless laws of impossibility." - The Rest Is Science

Logic at the heart of the task: P, Q, and two valid inferences

The hosts unpack the logic behind the conditional statement, introducing the concepts of the antecedent P and the consequent Q. They explain that the four-question card setup mirrors the possible truth values of P and Q, leaving four combinations to evaluate. They then discuss two valid forms of reasoning with conditionals: modus ponens (affirming the antecedent leads to the consequent) and modus tollens (denying the consequent leads to denying the antecedent). In the drinking example, modus tollens becomes a valid way to test the rule by looking for underage drinkers who are drinking alcohol and using that to falsify the rule.

"The antecedent is the really easy one where we say, here's a rule. If it rains tomorrow, Michael will stay indoors. It is now tomorrow and it is raining. Where is Michael? He's indoors. That's valid." - The Rest Is Science

Descriptive vs deontic framing, and social cognition

The conversation introduces a key distinction: descriptive rules describe what is, whereas deontic rules specify what ought to be done. The panel highlights that deontic framing (obligations and duties) often triggers a more intuitive compliance check, aligning with social norms. The duo references Lita Cosmidis and John Tuby who argued that human reasoning evolved more for social coordination than for universal logical deduction, coining terms like descriptive vs deontic to categorize puzzle versions.

"This is called a descriptive restriction. But when it comes to are people breaking the rule, we need to make sure people are complying with the rules. That's called a deontic rule." - The Rest Is Science

The 4% who get it right and what they reveal

The episode reviews historical data: original Wasson task success around 10%, replicated averages near 4%, and near-universal correct performance when the problem is framed around social rules. They discuss who tends to perform better—those with formal logic training, the mathematically oriented, or those who can connect the problem to real-world social norms. The dialogue sets up the question: is the task tapping a core logical competence or something more social and contextual?

"There was one paper that said the best correlation is whether you have taken logic classes, aha, yes, I am familiar with modus tollens and denying the antecedent." - The Rest Is Science

Counterexamples, counterexample hunting, and science ethics

A central thread is the distinction between seeking confirming evidence and actively looking for counterexamples. Johnson, Laird, and Watson’s classic 1970 study with two boxes of shapes illustrates how people focus on confirming evidence and ignore potential falsifiers until after it’s too late. The hosts tie this to Popper’s philosophy of science, praising Einstein for explicitly proposing counterexample tests to validate or falsify his theory, underscoring the value of counterexamples in progress.

"There's so much more information contained within a counterexample than there is in something that confirms it." - The Rest Is Science

Reasoning as a social tool and the path forward

The discussion concludes with a broad sociocultural view: reasoning likely evolved to help us cooperate and communicate, not to solve purely abstract puzzles. Mercier and Sperber’s Enigma of Reason is cited to illustrate that people often generate reasons after forming conclusions, using them to persuade others rather than to arrive at truth alone. The episode ends with reflections on public communication, the deficit model, and the need for deliberative, inclusive approaches to complex moral and scientific questions.

"We use reasons to explain ourselves and express who we are at a deep level. We don't use them because of anything logically timelessly true." - The Rest Is Science