Socratic questions: 6 types with examples
Socratic questions don't look for information. They look for the reasoning underneath it. They fall into six types: clarification, assumptions, reasons and evidence, perspective, implications, and the question about the question itself. Below is what each type does, when to bring it in, where it goes wrong — and wording you can take word for word.
An ordinary question asks what someone knows.
A Socratic question asks why they think it's true.What makes a question Socratic
Three things at once, and the third is the hard one.
The question is open — there's no answer you're already holding in mind. The question targets the reasoning, not the facts: not what happened, but what tells you that. And the question contains no hidden advice. That last one is the most common facilitator mistake: "Have you considered talking to her about it?" is grammatically a question and substantively a suggestion. The other person hears the suggestion.
The six-type breakdown comes from Richard Paul (Foundation for Critical Thinking), who developed it for teaching critical thinking. The Dutch tradition of the Socratic dialogue runs along a different line — Leonard Nelson, Gustav Heckmann, and in the Netherlands Jos Kessels and Het Nieuwe Trivium — and puts the emphasis on examining one concrete experience rather than an abstract claim. The six types work in both traditions; only the entry point differs.
In cognitive behavioural therapy, the same question form is called Socratic dialogue or guided discovery. The types below are recognisable there too, but the setting is one to one and the goal is to test an individual belief, not to get a group thinking together.
The six types of Socratic questions
1. Clarifying questions
They make what's been said more precise, before anyone forms a view about it. Skip this layer and the group spends the rest of the time discussing its own version of the story.
What exactly do you mean by "it isn't working"?
Can you name a specific moment when this happened?
What did you see happen, and what did you infer from that?
Who did what, in what order?
Are you saying it couldn't be done, or that it didn't happen?
Pitfall: the tone of an interrogation. Attach the question to the word, not to the person — "what does 'professional' mean here" lands differently from "what do you mean by professional". In the Socratic conversation, clarifying questions form a round of their own: the group only asks them after the case has been told, and the presenter answers factually, without prefacing anything.
2. Questions about assumptions
They surface the assumption sitting under a statement — the one the speaker no longer notices. This is the type that unsticks a stalled conversation fastest.
What has to be true for your conclusion to hold?
You say "I had no choice" — what made it unavoidable?
What are you assuming when you say she can't handle this?
Would someone who sees it differently be starting from a different premise?
Does that hold always, or only in this case?
Pitfall: the tone of exposure. Having an assumption laid bare feels like being caught out, unless you frame it explicitly as shared work: "Let's look at what's underneath this."
3. Questions about reasons and evidence
They separate what someone knows from what someone assumes. Bring them in as soon as the conversation consists of confident statements that nobody is checking the ground under.
What tells you that?
What here is observation and what is interpretation?
How do you know it was caused by this and not by that?
Is there anything from the same week that contradicts this?
What would change your mind?
Pitfall: "why" as an accusation. "Why did you do that?" triggers defence, even when it's meant neutrally. Replace it with "what tells you that" or "what led you to that". And if you want to discuss a case without any defence at all, the Gossip method is built structurally to do exactly that.
4. Questions about perspective
They put a second reading next to the first. Needed as soon as one account sounds so self-evident that nobody asks whether it could be otherwise.
How would the other person describe that same morning?
Who benefits from a different outcome here?
What would a colleague who often disagrees with you say about this?
Which reading haven't you taken seriously yet?
What would you advise if this happened to your colleague?
Pitfall: it turns into role play. The moment participants start playing "the other person", they defend a position instead of examining it. Keep it to one question and let the presenter answer.
5. Questions about implications and consequences
They follow the reasoning through: if this holds, what follows. The type that shows fastest whether a conclusion carries weight.
If this holds, what does it mean for next week?
What happens if everyone on the team does it this way?
What's the price of this solution, and who pays it?
What concretely changes if you do nothing?
What rule are you deriving from this — and does it still hold for the previous case?
Pitfall: sliding into action planning. The moment someone says "so what you should do is…", the inquiry is over. That's fine at the end and damaging in the middle.
6. Questions about the question
They examine whether the question the presenter walked in with is the question that needs solving. Paul counts this as a separate type, and in intervision it's the most underused one.
What's the question underneath your question?
Why does this question matter to you?
If you put it in one sentence, what sentence is that?
Are we answering your question right now, or a different one?
What would make an answer usable for you?
Pitfall: using this type only at the start. The return is greatest halfway through, when the group notices it's discussing something other than what it started with.
Cheat sheet: which type when
Six situations you'll hear in an intervision or a team conversation, and the question type that goes with each.
The group fills in what the presenter meant → clarifying questions, until everyone has the same picture.
Everyone agrees straight away → questions about assumptions; a shared assumption is an invisible assumption.
The conversation consists of opinions → questions about reasons and evidence.
One reading of the situation dominates → questions about perspective.
There's a solution on the table within five minutes → questions about implications.
The conversation drifts or stalls → questions about the question.
What you hear — and what to ask
What you hear Opening question "It just isn't working with that class." "What tells you it isn't working?" "I had no choice." "What made it unavoidable?" "She doesn't want this." "What did you see her do that you infer that from?" "That's just how it is around here." "Since when — and what happened before that?" "Then we should just set it up differently." "What's the price of that, and who pays it?" "What were we talking about again?" "What question are we actually answering right now?"
The four mistakes that make a Socratic question harmless
Advice in disguise
"Shouldn't you go and talk to her?" Anything starting with shouldn't you is a suggestion in the shape of a question.
Stacked questions
Three questions in one breath. The other person answers the easiest one and the other two are gone. Ask one question, then stay quiet.
"Why" as an accusation
See type 3. Same content, different opening, different conversation.
Going abstract too early
"What is fairness, really?" is a good question — after the case has become concrete, not before. The Nelson tradition is strict about this: the general emerges from the concrete, not the other way round.
When Socratic questions don't fit
Probing further isn't always the right move. Four situations where you'd better not:
In acute danger or a crisis — that calls for instruction, not inquiry.
When factual information is missing that you could simply look up.
When a decision has to be made within ten minutes.
When the presenter is emotionally overloaded; support comes first.
Watch the power relation too. A manager who consistently probes an employee's assumptions isn't running a Socratic conversation but an appraisal — unless the role division has been agreed in advance and stays visible.
The kind of case where these questions come into their own: a recurring pattern, a disagreement that isn't about facts, a situation that has been on the agenda three times already. If it's one bounded incident that needs to be factually unpicked, the Incident method is the more precise instrument.
How to put these questions into a conversation
Individually good questions won't save a conversation — the sequence does the work. Clarification first, then assumptions and evidence, then perspective, implications only at the end. That's exactly why the Socratic conversation has fixed steps with a time window of their own: the timer runs alongside you, so you don't have to watch the clock while facilitating.
If your group works better with something tangible, put your own questions on cards and print them — eight per A4 sheet via the card generator. One type per card works well: participants holding an "assumptions" card ask noticeably fewer advice-questions.
A good question moves nothing in the room.
It moves something in the other person's head.
Sources
Leonard Nelson, Die sokratische Methode (1922) — the origin of the Socratic dialogue as a group form.
Gustav Heckmann — on the facilitator's role and suspending your own judgement.
Jos Kessels, Socratische gesprekken; Het Nieuwe Trivium — the Dutch elaboration for organisations and intervision.
Richard Paul & Linda Elder, The Thinker's Guide to the Art of Socratic Questioning, Foundation for Critical Thinking — the six-type breakdown.