From Socrates as Plato portrays him (Meno and Republic, c. 380 BCE), set out for modern use by Richard Paul and Linda Elder (2006).
Tests a belief through layered questions about meaning, evidence, assumptions, and consequences, so the person holding it finds the gaps in their own reasoning.
One confident claim, tested by four kinds of question
A manager says remote staff get less done. At each of the questioner's turns, pick a question about meaning, evidence, assumptions, or consequences, or the counterargument that tells him the answer. His reply shows whether he is examining the claim or defending it. Illustrative.
HR partner
Manager
HR partner
Manager
Questions that let him find it: 0 of 4Questions that let him find it: 1 of 4Questions that let him find it: 2 of 4Questions that let him find it: 3 of 4Questions that let him find it: 4 of 4
- The claim (Manager). "Remote people just get less done. We should bring everyone back five days a week."
- Meaning (HR partner). Pick the first question.
- That's not what the research on remote work says. I don't care what some study says. I know my team. A counterargument: he defends the claim, and now it is his position.
- When you say less done, what would you measure? Claims closed, I suppose. And how fast they answer me. A question of meaning: the claim now has two parts he can check.
- A definition (Manager). Two measures he chose himself. One is output; the other is something else.
- Evidence (HR partner). Pick the question.
- What have your own team's claims numbers shown since people went remote? Honestly, I haven't looked. It's more that I can't see them working. A question of evidence: he finds that the claim rests on no data.
- Your remote people actually close more claims than the office ones. Then the numbers are wrong, or they're closing the easy ones. Telling him: a fact he did not find is a fact he can dismiss.
- A gap (Manager). The claim was about output; what he has is a feeling about visibility.
- Assumption (HR partner). Pick the question.
- So it's really about control, isn't it? That's unfair. I'm trying to run a team. A verdict on his motive: he defends his character.
- Is the worry that output is lower, or that you can't see the work happening? The second, I guess. When something's stuck I find out late. A question of assumptions: he separates the two and names the real one.
- The real worry (Manager). The problem has moved from where people sit to how fast blockers surface.
- Consequences (HR partner). Pick the last question.
- Then just ask everyone for a daily status update. Fine, I'll do that. A conclusion handed over: he takes it, and it is not his.
- If everyone came back five days, would you hear about stuck work sooner? Not really. People hide stuck work in the office too. What I need is to hear about blockers early. A question of consequences: the conclusion is his own.
- His conclusion (Manager). He drops the five-day plan and asks for a ten-minute check on blocked claims each morning.
An example
An operations manager at an insurance company told his HR partner he wanted his whole claims team back in the office five days a week, because remote people simply got less done. She had the team's numbers and could have shown him on the spot that the remote adjusters closed slightly more claims.
She asked questions instead. What would less done look like, if he measured it? Claims closed, he said, and how fast people answered him. What had the claims numbers shown since the team went remote? He had not looked; his sense came from not being able to see anyone working. Was the worry that output was lower, or that he could not see the work? The second, he said: when a claim was stuck, he found out late. Would five days in the office surface stuck claims sooner?
He thought about it and said no, people hid stuck work at their desks too. He dropped the mandate and set up a ten-minute morning check on blocked claims.
Why it works
A belief someone holds with confidence is rarely changed by counterevidence. Presented with facts that contradict it, people defend the belief, question the facts, and leave more certain than before. The method avoids that contest. Each question asks the holder to look at one part of their own reasoning, what the claim means, what it rests on, what it assumes, and what follows from it, and the holder does the examining.
The gap the holder finds is one they cannot dismiss, because they found it in their own answers. That is also why the conclusion lasts. The manager in the example would have argued with a chart; he could not argue with his own admission that he had never looked at the numbers. The questioner's restraint is the method: every fact she holds back is one the other person gets to discover.
The thinking behind it
Plato Plato's dialogues, among them the Meno and the Republic, are the main record of the method of questioning he gave to Socrates: ask a person to define a term, then draw out what follows from the answer until its contradictions show. The dialogues are Plato's portrayal of Socrates, not a transcript of the historical man, and the method on this page is a business use of what he depicted.
Richard Paul Richard Paul, a philosopher who founded the Center for Critical Thinking, wrote The Thinker's Guide to the Art of Socratic Questioning (2006) with Linda Elder, president of the Foundation for Critical Thinking. The guide sits within their framework of the elements of thought and intellectual standards, and includes a taxonomy of Socratic questions, the kinds of question a questioner can ask to probe meaning, evidence, assumptions, and implications.
How to run it
- State the claim. Say the belief back in one sentence and get agreement that this is what is being examined.
- Ask what it means. Ask for a definition or a measure: what would less productive look like, counted?
- Ask what it rests on. Ask how they know, and what they have seen themselves.
- Ask what it assumes. Look for the belief under the belief, and offer two readings to choose between.
- Ask what follows. Ask what would happen if the claim were true and acted on, and who would see it differently.
- Hold back your answer. Keep your own evidence in reserve, since a fact you hand over is a fact they can dismiss.
- Let them conclude. Stop asking once they have named the gap, and let the conclusion be theirs.
Where it goes wrong
The questions turn into a cross-examination. A run of questions that each corner the other person feels like a trap, and people who feel trapped stop answering honestly. Ask from curiosity, and accept answers that surprise you.
Leading questions smuggle in the answer. "Isn't it really about control?" is an accusation with a question mark, and it puts the holder on the defensive. If a question has only one acceptable answer, it is a statement.
The method is used where it does not belong. When a decision is urgent or the other person is your manager in a crisis, a slow walk through their reasoning reads as stalling. Save it for beliefs that will shape decisions over time.
When to use something else
- Calibrated Questions. When the aim is to win room in a negotiation, not to test a belief's footing.
- Steelmanning. When the claim deserves its strongest form before anyone tests it, especially one you disagree with.
- Toulmin Model. When an argument is written down and its claim, grounds, and warrant can be laid out and checked.
Where it sits in SOLVE eX
- Any step: Tools that serve every step: habits of thought and checks on the thinking itself.
Go to the source
- Meno; Republic, c. 380 BCE
- The Thinker's Guide to the Art of Socratic Questioning, 2006