The Socratic Method: Why Questions Teach Differently Than Answers

Something is happening to the Socratic method right now that hasn't happened in a while: people outside philosophy departments and law schools are arguing about it. Khan Academy built its AI tutor, Khanmigo, around the explicit promise that it will not just hand students an answer — it will ask the question that gets them there themselves. Teachers worried about students outsourcing their thinking to chatbots keep reaching for the same phrase to describe what they want instead: more Socratic questioning, less answer-retrieval. Even the debates about AI and critical thinking keep circling back to a 2,400-year-old teaching method as the thing we're afraid we're losing.

I think that's worth pausing on, because most people invoking "the Socratic method" right now have a fuzzy idea of what it actually was. It gets used as a synonym for "asking a lot of questions" or "being hard on students," which flattens something much sharper and, in my view, much more useful.

What the Socratic Method Actually Is

Everything we know about Socrates comes secondhand, mostly through Plato's dialogues, since Socrates himself wrote nothing down. In those dialogues, he almost never lectures. He asks someone to define a term they've been using confidently — justice, courage, piety — and then asks follow-up questions until the definition collapses under its own contradictions. The Greek word for this is elenchus, which translates closer to "cross-examination" or "refutation" than to "teaching," and that distinction matters. Socrates wasn't trying to transmit knowledge he already had. He was trying to expose the gap between what someone thought they knew and what they could actually defend.

The clearest example is the Meno, where Socrates walks an enslaved boy with no formal education through a geometry problem — doubling the area of a square — using nothing but questions. The boy gets it wrong first, sees why, and then gets it right. Socrates never states the answer. He treats being wrong as the necessary middle step, not a failure to route around.

That's the part of the method most modern uses skip. The goal wasn't the destination. It was the productive confusion along the way, a state the Greeks called aporia — genuinely not knowing, together, in a way that opens the door to actual learning instead of borrowed certainty.

The Line Everyone Quotes, and What It Actually Meant

Plato's Apology records Socrates' defense at his trial, and it contains the line that outlives everything else he's credited with: "the unexamined life is not worth living for a human being" (Apology 38a). People quote it as inspirational wallpaper. In context, it's closer to a threat than a slogan. Socrates is telling the jury that is about to sentence him to death that he will not stop questioning people in the street, because a life spent unexamined isn't one worth protecting through silence. The line is a justification for the method, not a decoration on top of it. He wasn't asking people to reflect more. He was saying that a life without that kind of scrutiny doesn't meet the bar of being fully lived.

I find that a much harder claim to sit with than the inspirational-poster version. It says the questioning wasn't optional or occasional. It was the whole point.

Why It's Trending Again

The Khanmigo example is the clearest current data point. Khan Academy's own framing of the tool, since its 2023 launch, has centered on the idea that it should guide a student toward an answer through questions rather than simply supplying one — an explicit, named attempt to encode Socratic method into a piece of software used by students at scale. That's a genuinely new thing: a method built for a person standing in a public square is now being built into a chat interface.

The same instinct shows up in the classroom debates about AI and critical thinking. When a student can get a finished essay or a solved problem in ten seconds, the value of a teaching method that refuses to give finished answers goes up, not down.

Socratic seminars — structured classroom discussions where students question a text and each other rather than receive a lecture on it — have been a fixture of humanities classrooms for decades, but they're getting reframed lately as a countermeasure specifically against answer-dependence.

There's also a much older, quieter home for the method: American law schools have used a formal version of it, often called the case method, since Christopher Columbus Langdell introduced it at Harvard Law School in the 1870s. First-year law students still get cold-called and cross-examined about a case's reasoning rather than lectured about the rule it produced. That tradition never went away. It's just gotten more visible as a reference point now that people are looking for models of rigorous, answer-resistant dialogue.

What connects all three — the AI tutor, the classroom seminar, the law school cold-call — is the same underlying bet: that being made to construct an answer under sustained questioning teaches something that receiving the answer cannot.

The Method Versus the Performance

Here's where I want to be honest about the method's reputation problem. Ask most adults what "Socratic method" feels like and they won't describe curiosity. They'll describe being cornered. A professor picks you out, asks a question, picks apart your answer, asks another question that makes your first answer look worse, and keeps going until you're visibly stuck in front of everyone. That's not a caricature — it's a legitimate description of how the method gets practiced badly, and it happens often enough that the method has earned a reputation as intellectual hazing dressed up in Greek.

The distinction I keep coming back to is that elenchus was never supposed to have a winner. Socrates included himself in the confusion. He claimed, famously, to know that he knew nothing — and whether or not that was fully sincere, it set a tone. The questioner was supposed to be exposed alongside the person being questioned, both of them working toward something neither had already secured. A cold-call that exists to make a student feel small has kept the interrogation and dropped the shared vulnerability. That's not a lesser version of the Socratic method. It's a different thing wearing its name.

Socratic Method Compared to Other Dialogue Formats

Laid out this way, the Socratic method sits closer to reflective dialogue than to debate, even though its reputation suggests the opposite. Debate assumes both sides already know what they think and are trying to win. The Socratic method assumes the opposite: that the person being questioned doesn't yet know what they think as clearly as they believe they do, and the questioner doesn't either. That's a much more vulnerable posture than a debate stage allows, which is probably why so few people practice it in its original form. It's a lot easier to be right than to be genuinely uncertain in front of someone else.

What Made It Work in the First Place

I think three things had to be true for elenchus to function the way Plato describes it, and losing any one of them is usually what breaks a modern attempt to use it.

1. The questioner had to be asking because they didn't know — not because they already had the answer loaded and were waiting to reveal it. A question used as a trap isn't a Socratic question. It's a rhetorical one wearing a question mark.
2. The respondent had to be willing to follow their own answer to an uncomfortable place. Plato's dialogues are full of people who get cornered and simply refuse to keep going — Euthyphro walks away mid-conversation rather than sit with the contradiction Socrates has surfaced. The method only works on someone who stays in the room.
3. It took time. Aporia isn't reached in one exchange. It takes several rounds of question and answer, each one narrowing the gap between what someone said and what they can defend, before the contradiction becomes visible enough to sit with. A single gotcha question isn't the method. It's a fragment of it, mistaken for the whole.

That third requirement is probably the biggest reason the method degrades so easily in modern settings. A classroom period, a legal argument, a chatbot exchange — none of them naturally protect the time it takes to get past the first uncomfortable answer and into the second or third. Skip the sustained back-and-forth and you're left with the cornering feeling and none of the shared discovery that was supposed to justify it.

I've written elsewhere about a related practice we use for testing ideas without testing relationships, and the design problem is the same one Socrates ran into: how do you keep sustained, uncomfortable questioning from curdling into a contest of who can make whom look foolish first. The answer isn't to soften the questions. It's to build enough structure and enough goodwill into the room that both people can stay in it past the point where the first answer stops holding up. My piece on how the Socratic hotseat teaches intellectual humility goes further into what that structure actually looks like in a group setting, and how it differs from the classroom cold-call most people picture when they hear the word "Socratic."

Bringing the Method Into a Room, Not Just a Transcript

Plato's dialogues are almost always one-on-one. Socrates and one other person, working through a single claim until it either holds up or doesn't. Most modern uses of the method — the classroom, the law school, the AI tutor — try to preserve that one-on-one intensity while an audience watches, which is exactly the setup that produces the cornering feeling I described above. The audience turns a shared search into a performance, and performances have winners.

What I've found works better in group settings is treating the questioning itself as a rotating, shared job rather than something one authority figure does to one respondent while everyone else watches. When the group takes turns being the questioner and the questioned, the aporia becomes something the room reaches together instead of something inflicted on one person at a time. It's still uncomfortable. It should be. But it stops being a spectator sport, which is the thing that turns rigorous questioning into humiliation. My piece on the tribunal format for subjecting ideas to rigorous, respectful scrutiny walks through one version of what that rotation can look like.

Where It Breaks Down, and How to Guard Against It

The most common failure mode isn't cruelty. It's asymmetry. One person holds more authority, more confidence, or more practice with the format, and the questions stop being a shared search and start being a demonstration of who's smarter. Socrates himself wasn't fully exempt from this criticism — plenty of his interlocutors in Plato's dialogues end up looking foolish while Socrates ends up looking wise, and it's fair to wonder how much of that is the method working and how much is Plato writing his teacher as the hero of his own dialogues.

The honest fix isn't to abandon sustained questioning. It's to build in the thing Socrates claimed for himself and rarely extended visibly to others: real uncertainty on the questioner's side too. A Socratic exchange where only one person risks looking wrong isn't really Socratic. It's cross-examination with better branding.

FAQ

What is the Socratic method in simple terms? It's a method of teaching or inquiry where someone asks a series of questions about a belief a person holds, rather than lecturing them, until the person sees a gap or contradiction in their own reasoning. It comes from how Plato depicts Socrates arguing in dialogues like the Apology and the Meno.

Is the Socratic method the same as debate? No. Debate assumes both sides already know their position and are trying to win. The Socratic method assumes the person being questioned — and ideally the questioner too — doesn't yet know their position as clearly as they think, and the goal is to find that out together rather than to score a win.

Why are AI tutors using the Socratic method now? Tools like Khan Academy's Khanmigo, launched in 2023, are built to guide students toward answers through questions instead of simply supplying the answer, an explicit attempt to encode Socratic-style questioning into software at a moment when easy answers from AI are a widely shared concern in education.

Do law schools really still use the Socratic method? Many do, in the form of the case method introduced at Harvard Law School by Christopher Columbus Langdell in the 1870s, where students are questioned about a case's reasoning in class rather than given the rule directly.

Why does the Socratic method sometimes feel like an attack instead of a conversation? Because the sustained-questioning part of the method survives easily while the shared uncertainty part doesn't. When only one person in the exchange risks looking wrong, cross-examination replaces the mutual search for understanding that the method depended on in Plato's dialogues.

Last updated: 2026-08-22