Check Your Reasoning with the Socratic Method

What Is the Socratic Method?

Choose [Socratic Method] from the mode row
              ↓
┌────────────────────────────────────────────────┐
│ Socratic Method    [× Exit]                    │
│ ────────────────────────────────────────────── │
│ I think … because …  [Tier icon] [Send]        │
└────────────────────────────────────────────────┘

The selected mode and any progress information appear at the top of the white message box, above a thin divider. If there is no conversation history, you will see icons, brief instructions, and an example. Select About this mode to ask the AI for an explanation. To switch modes, select × and choose another one.

You may have an opinion before you have fully worked out why you hold it. Say what you think, then ask yourself why. The Socratic Method starts with that idea: the AI follows up on one important point to help you notice anything you have left out.

A follow-up looks at a particular part of the reasoning. After thinking it through, you may find that your judgment still has support. You can keep it. Or you may need to add a condition or revise the conclusion as the discussion develops.

If you only want to check a fact or get instructions, Q&A is a more direct choice.

State Your View and Explain Why

Suppose you are studying plant care. You might say:

I think the longer a plant stays in the sun, the better it grows, because plants need sunlight.

That idea is enough to start the discussion. If you are unsure, you can also say “I'm unsure which of these two options to choose,” and explain why you are hesitating.

The AI builds on what you said and asks one key question at a time, such as:

Do all plants thrive in strong sunlight for long periods?

This question returns to your original judgment: do different plants need the same amount of light? That condition can change your conclusion.

Refine Your View

You might realize:

Plants need light, but different plants need different amounts. I'll find out what this plant needs before deciding where to put it.

Your original reason still holds: plants need light. You have also added a condition: consider the needs of the specific plant.

The next question may check something different: what a term means, what supports the view, whether another explanation is possible, or whether the judgment holds in a different situation. Take the question in front of you first. Work through that before turning to the next point.

State a judgment and reason → Answer one follow-up question
                                          ↓
Add or revise → Reach a conclusion the discussion currently supports

Once the reasoning is clear, you can finish the round. Sometimes the discussion shows you which information you still need. You can pause there: “I still need to know … before deciding.”

Understand the Follow-Up

Suppose you say “Yellow leaves mean the plant needs water.” The AI might ask:

What else could cause leaves to turn yellow?

Too little water is one possibility. You could also check whether the plant has been overwatered or is getting the right amount of light before deciding.

If the AI makes your view more absolute than you intended, such as turning “This fits the current situation” into “This fits every situation,” correct it directly:

I'm only talking about this situation. Please keep the discussion within that scope.

Check that the discussion is still about what you meant. If the scope has drifted, bring it back. If you can explain and support your original view, you can keep it.

When You Need Help

Say “Rephrase the question,” “Narrow the scope,” or “Give me an example.” The AI will help with the current question.

If you lack a basic concept, switch to Q&A manually, clear it up, and return. Q&A is also more direct when you want a full answer or an organized explanation.

The Feynman Technique focuses on explaining the content clearly. The Socratic Method focuses on whether your judgment and reasons make sense. Combine them in the order that suits your current need.

Start and Stop

Choose Socratic Method and send “Start” for the AI to suggest a question from the chapter, or state your own judgment directly.

Say “Continue” to resume the current line of questioning. Specify a new question if you want to change it, or say “Pause” or “Stop” to stop.

The round ends once the question is clear. You choose what to discuss next and when to switch modes or chapters.

Use the Feynman Technique to turn your reasoning into a clear explanation. If you are unsure what to work on next, read learning status guidance.