Socratic Questioning: 6 Types of Questions, With 60 Examples

Maya teaches ninth-grade history. One of her students, Jonas, announces that the printing press was "basically the internet of the 1400s." Maya does not correct him. She asks: "What do you mean by basically?" Then: "What would have to be true for that comparison to hold?" Then: "Who could use a printing press in 1480, and who could use the internet in 2010?"

Four questions later, Jonas is the one saying the comparison works for speed of spreading ideas but falls apart on who got to publish. Nobody told him he was wrong. He found the edges of his own claim, so he remembers them.

The same thing happens in offices. Elif, an engineering manager, has a developer who wants to rewrite a service from scratch. Instead of saying no, she asks what problem the rewrite solves, what they would lose, and how they would know in three months whether it worked. The developer ends up proposing a smaller refactor on his own.

That is Socratic questioning: changing how someone thinks, including yourself, by asking rather than telling. This guide gives you the six classic types of Socratic questions, ten example questions for each (sixty in total), a short dialogue for each type, and advice on using them in class, in coaching, at work and in your own journal.

What Socratic questioning is

Where it comes from

Socrates, the Athenian philosopher of the fifth century BC, wrote nothing down. We know his method mostly through the dialogues of his student Plato (and, less extensively, through Xenophon). In them, Socrates asks people what courage, justice or piety is. They answer confidently, and a few questions later their definition contradicts something else they believe.

The Greek name for this cross-examination is elenchus. Its typical shape: someone states a belief, Socrates gets them to agree to a few other statements, and then shows that those statements clash with the original belief. The person is left in aporia, a productive state of puzzlement where they realize they do not know what they thought they knew.

You will often see the line "I know that I know nothing" attributed to Socrates. It is a popular paraphrase, not a direct quote. In Plato's Apology, Socrates says something closer to this: unlike the people he questioned, he does not think he knows what he does not know. The difference matters. Socratic humility is not claiming total ignorance. It is being honest about the limits of what you know.

How it differs from interrogation

A police interrogation and a Socratic conversation both involve lots of questions. The difference is intent.

Modern Socratic questioning is gentler than Plato's Socrates, who could be relentless. The aim is not to humiliate anyone into aporia but to help them reach a clearer, better-supported view, which might be the view they started with.

The six types of Socratic questions

The most widely used taxonomy of Socratic questions sorts them into six types. It is commonly attributed to Richard Paul, a critical thinking scholar whose work with the Foundation for Critical Thinking popularized it among teachers.

You do not need to use them in order. In practice, a good conversation loops: clarify, probe a reason, clarify again, look at a consequence. But the list is useful as a menu when you get stuck.

1. Clarification questions

Purpose: Most disagreements start with fuzzy words. Clarification questions make a claim specific enough to examine. They are also the least threatening type, so they are a good place to start.

10 examples:

1. What do you mean by that?
2. Could you put that another way?
3. Can you give me an example?
4. What is your main point here?
5. How does this relate to what we were discussing?
6. When you say "always," do you mean always, or usually?
7. What would a counterexample look like?
8. Is your point that X, or that Y?
9. Can you say more about that?
10. How would you explain this to someone who has never heard of it?

Mini dialogue:

Jonas: Social media is bad for teenagers.

Maya: Bad in what way?

Jonas: It makes them anxious.

Maya: All social media, or some kinds more than others?

Jonas: Mostly the ones where you post pictures and count likes, I guess.

Jonas's claim just went from vague and sweeping to specific and testable.

2. Questions that probe assumptions

Purpose: Every argument rests on things the speaker takes for granted. These questions bring those assumptions into the open, where they can be checked. This is often where the biggest shifts in thinking happen.

10 examples:

1. What are you assuming here?
2. What could we assume instead?
3. Why do you think that assumption holds?
4. Does this always hold, or only in certain situations?
5. What would change if that assumption were false?
6. Where did this belief come from?
7. Is that something you checked, or something you have always accepted?
8. You seem to be assuming X. Is that right?
9. Would everyone in this situation share that assumption?
10. What would you have to believe for the opposite to be true?

Mini dialogue:

Elif: You said we should hire a senior engineer instead of two juniors. What is that based on?

Developer: Seniors ship faster.

Elif: You are assuming speed is our main bottleneck right now. Is it?

Developer: Actually, our bigger problem is that nobody has time to write tests.

Elif: Would a senior or two juniors help more with that?

3. Questions that probe reasons and evidence

Purpose: Once a claim is clear and its assumptions are visible, you check what supports it. These questions ask for reasons, test their quality, and ask what would change the person's mind.

10 examples:

1. How do you know that?
2. What is your evidence?
3. How reliable is that source?
4. Is that enough evidence to be sure?
5. Can you give an example that supports this?
6. What would change your mind?
7. Could there be another explanation for the same facts?
8. Are these reasons good enough, or just the first ones that came to mind?
9. Is this based on one experience, or many?
10. How could we find out whether that is true?

Mini dialogue:

Maya (to herself): I am terrible at public speaking.

Maya: How do I know that?

Maya: My presentation last spring went badly.

Maya: One presentation? How did the other three go this year?

Maya: Fine, honestly. One was good.

Notice question 6 on the list. "What would change your mind?" is one of the most powerful questions you can ask anyone, including yourself. If the answer is "nothing," it is a belief, not a conclusion.

4. Questions about viewpoints and perspectives

Purpose: Most views look strongest from the inside. These questions step outside, asking how the issue looks to someone with different interests, values or information. Done well, this is close to steelmanning, which you can read more about in our guide on steelmanning.

10 examples:

1. How would someone who disagrees respond?
2. What is the strongest argument against your view?
3. Who benefits from this, and who pays?
4. How might this look to someone from a different generation or background?
5. What would [a specific person] say about this?
6. Is there another way to look at it?
7. What are the strengths and weaknesses of each view?
8. Why might a reasonable person believe the opposite?
9. How would you see this if you were on the receiving end?
10. What do both sides agree on?

Mini dialogue:

Jonas: School uniforms are pointless.

Maya: Why might a parent want them?

Jonas: Maybe because they are cheaper than buying trendy clothes?

Maya: And a student who cannot afford trendy clothes?

Jonas: Okay, they might like them. But they still kill self-expression.

Maya: So the real tradeoff is cost and pressure versus self-expression?

5. Questions about implications and consequences

Purpose: Ideas have consequences. These questions follow a claim forward: if it is true, what else must be true, and what happens if we act on it? They are especially useful for decisions and plans.

10 examples:

1. If that is true, what follows?
2. What are the consequences of that assumption?
3. How would this affect other people?
4. What happens in a year if we do this? In five years?
5. What is the best case and the worst case?
6. Does this fit with what you said earlier?
7. If everyone did this, what would happen?
8. What would we have to give up?
9. How would we know if it was working?
10. What is the smallest step we could take to test this?

Mini dialogue:

Developer: We should rewrite the billing service in a new language.

Elif: Suppose we do. What happens to feature work during the rewrite?

Developer: It would pause. Maybe three months.

Elif: And what does sales tell customers during those three months?

Developer: Hmm. What if we only rewrite the invoicing module first?

6. Questions about the question

Purpose: Sometimes the problem is the question itself. It may be the wrong question, a loaded one, or one that cannot be answered as asked. These meta questions step back and examine the inquiry.

10 examples:

1. Why is this question important?
2. Is this the right question to ask?
3. What is this question assuming?
4. Could we break this question into smaller ones?
5. What would a good answer look like?
6. Is this question answerable with the information we have?
7. Why do you think I asked that?
8. Is there a better question we should be asking?
9. What did we learn from trying to answer this?
10. Who gets to decide what counts as an answer here?

Mini dialogue:

Jonas: Who was the best president ever?

Maya: Best at what?

Jonas: Just, overall.

Maya: What would we need to agree on before we could answer that?

Jonas: What "best" means. Wartime leadership, economy, civil rights...

Maya: So maybe the better question is: which president did the most for one of those, and why?

Using Socratic questioning in the classroom

The Socratic seminar is the classic classroom format: students sit in a circle, discuss a text or a question, and the teacher mostly asks questions rather than lecturing. A few practices make it work.

• Choose a question with real room for disagreement. "What year did the war start?" is a fact. "Was the war avoidable?" is a discussion.
• Use wait time. After a question, count silently to five before you rephrase or answer it yourself.
• Question the answer, not the student. "What makes you say that?" works for correct and incorrect answers alike, so students cannot tell from your question whether they were right.
• Let students question each other. Post the six types on the wall and ask students to pick one to use on a classmate.
• Close with type 6. "What question are we leaving with?" is a strong end to any seminar.

For self-study, an AI study buddy can play the questioner when there is no teacher in the room, quizzing you on why rather than what.

Coaching and therapy-adjacent contexts

Coaches use Socratic questioning because people are more committed to solutions they reach themselves. A good coaching question ("What would you advise a friend in your situation?") does the same work as type 4, and "What is one thing you could try this week?" is type 5 in action.

Socratic questioning is also part of cognitive behavioral therapy (CBT), the approach associated with Aaron Beck, where therapists use guided questions to help clients examine the evidence for and against their thoughts. The boundary is clear: using Socratic questions on yourself or with friends is a thinking habit, not therapy. If you are dealing with persistent anxiety, depression or anything that affects your daily life, talk to a qualified professional. A list of questions is not a substitute.

Socratic questioning at work

At work, Socratic questioning replaces "here is what you should do" with questions that build someone's judgment.

In 1:1s

Elif's go-to 1:1 questions:

• "What is the hardest thing you are working on right now, and why is it hard?"
• "What have you already tried?"
• "What would you do if I were not available this week?"
• "What would success look like by Friday?"

The last two build independence.

In code reviews

Questions in code reviews come across as less combative than directives, and they sometimes reveal that the reviewer was the one missing context.

• "What happens here if the list is empty?" (type 5)
• "Why did you choose a cache over a database query?" (type 3)
• "Is this assuming requests arrive in order?" (type 2)

One warning: a question that is really a command ("Don't you think this should be a constant?") is just a directive wearing a disguise. If you know what you want, say it plainly.

In product decisions

Before a big decision, run the six types as a checklist. What exactly are we deciding? What are we assuming about users? What evidence do we have? How would support, sales and a skeptical customer see it? What happens in six months? And is this even the right decision to be making now? For a longer process, see how to make better decisions with AI.

Here is a question product teams argue about constantly:

Socratic questioning on yourself

You can run the whole method on your own beliefs. The easiest format is a journal. Pick a belief that is bothering you, write it at the top of a page, and answer these prompts in order:

1. What exactly do I believe? Write it in one sentence.
2. What do I mean by the key word in that sentence?
3. What am I assuming for this to be true?
4. What is my evidence for it? What is my evidence against it?
5. How would someone who cares about me, but disagrees, see this?
6. If this belief is true, what follows? If it is false, what follows?
7. What would change my mind?
8. Is this the right question to be worrying about?

The prompts are the six types in order, plus "what would change my mind" as a pivot. You will not always change your view. The goal is to hold it for reasons you can name.

Common mistakes

Leading questions. "Wouldn't you agree that remote work hurts collaboration?" is a statement with a question mark at the end. People notice, and they stop thinking and start defending. Ask open questions: "How has remote work affected collaboration on your team?"

Gotcha questions. Setting a trap so you can say "aha, so you contradict yourself" wins the moment and loses the person. Plato's Socrates sometimes did exactly this, and in the dialogues people often leave irritated. If you catch a contradiction, name it as a shared puzzle: "These two things seem to pull against each other. How do they fit?"

Too many questions at once. "What do you mean, and why do you think that, and what would your boss say?" is three questions, and the person will answer the easiest one. Ask one. Wait. Listen.

Not listening to the answer. If your next question was ready before they finished speaking, you are running a script, not a conversation. The best follow-up question almost always comes from a specific word in their answer.

Hiding behind questions. If someone asks what you think, tell them. And someone who is grieving or furious needs to be heard, not cross-examined.

If you want to spot the reasoning errors these questions tend to uncover, our guide to logical fallacies is a good companion.

A worked example: "I am too old to change careers"

Elif's friend Maya is 41 and has taught history for fifteen years. Over coffee, she says she would love to become a UX designer, but she is too old. Here is how Elif, using Socratic questions, helps her examine that belief. The question type is in brackets.

Elif: What do you mean by "too old"? [clarification]

Maya: Companies want young designers. And it takes years to get good.

Elif: So there are two claims: employers will not hire you, and you do not have enough time to get good. Is that right? [clarification]

Maya: Yes, that is it.

Elif: Let's take the first one. How do you know employers do not hire career changers in their forties? [evidence]

Maya: I don't, really. It is just a feeling. Tech feels young.

Elif: Do you know anyone who switched into design or tech later? [evidence]

Maya: Actually, one of the designers at my cousin's company used to be a nurse. She is probably 45.

Elif: So "they never hire older people" might be more like "it may be harder"? [assumptions]

Maya: Fair.

Elif: Now the second claim. You are assuming you would be starting from zero. Are you? [assumptions]

Maya: Well, I use education software every day. I have complained about bad gradebook and classroom app interfaces for fifteen years.

Elif: How would a company building education software see that? [viewpoints]

Maya: Hmm. As domain knowledge, probably. Maybe even an advantage.

Elif: Suppose you started learning this month. Where would you be at 43? [implications]

Maya: A designer with two years of practice and fifteen years in classrooms.

Elif: And if you do not start, where will you be at 43? [implications]

Maya: A teacher who still wishes she was a designer.

Elif: So maybe the question is not "am I too old?" What is the real question? [question about the question]

Maya: Whether I want it enough to spend two years of evenings on it. And whether I can test that before quitting.

Elif never said "you are not too old." She did not give advice. Maya walked from a flat "I can't" to a concrete, testable question she owns.

Want to feel this from the receiving end? Try a belief of your own with a questioner:

Practicing Socratic questioning with AI

Getting comfortable asking these questions, or being asked them, takes practice. An AI partner will question you at midnight and never get tired or offended.

On Debatly, there are a few ways to practice:

• Chat with Socrates. Bring him a belief and he will ask about definitions and hunt for contradictions, much like in Plato's dialogues.
• Chat with the Socratic questioner. A modern, coaching-style questioner, good for a decision or plan you want to stress-test.
• Run a Socratic debate. In the Socratic format, you get a questioner instead of an opponent. Instead of arguing back, the AI keeps asking until the weak points of your position show up. You can start one from the debate room, and our Socratic method AI page explains how it works in more detail.

If you prefer a sharper opponent, a devil's advocate argues against you directly, which pairs well with Socratic practice: first let a questioner find your assumptions, then let a critic attack them.

An honest limit: an AI questioner does not know your life, and it can ask questions that miss what really matters to you. Use it to practice the technique and to find blind spots in arguments, and bring the conversations that really matter to people you trust.

Start asking better questions

Pick one type this week and use it deliberately. Clarification is the easiest start: the next time someone makes a sweeping claim, ask "What do you mean by that?" and actually listen to the answer.

If you want a practice partner, open a Socratic debate on Debatly with a belief you care about, and see which question makes you pause. Try for free.