Thesis defence

Defend a claim under sustained, specific questioning from someone who is trying to break it, and find out which parts of your position are load-bearing and which were ornamental.

When to use this: When checking yourself
Learning impact: Strong change
Learning evidence: Credible with caveats
Where you are with the topic: Checking readiness
How often to use it: Worth repeating often

When to use it

An essay, viva, or seminar needs a claim that survives sceptical questioning.
You feel fluent after reading, then blank when someone asks you to explain without notes.
You finished drafting a thesis sentence and want to know what would break it.

What to ask AI

Type something like: “I'm defending this thesis: [claim]. Act as a sceptical examiner in my field.”

Require one objection at a time, and ask it to remember what you said earlier.

Answer each objection before you ask for the next.

How it helps

Thesis defence is sustained, specific questioning of a claim you are committed to.

A friendly friend lets you drift. A good examiner remembers what you said four turns ago.

Thesis drift is quiet: you weaken the claim to absorb each counterexample, then present the softer claim as if it were the original.

From the inside that feels like refining. From the outside it is retreat.

Defending under pressure forces you to name premises, evidence, and what would count as disconfirmation.

Material used under pressure sticks better than material only reviewed.

How to run it

Commit to a real thesis, a claim someone could disagree with. “This topic is complex” cannot be defended.
Ask for a sceptical examiner who tracks your wording across turns.
Answer; do not hedge. Qualification that quietly changes the claim teaches nothing.
When you concede, say so out loud. That prevents invisible drift.
Go until something load-bearing breaks, or until you can name the premise you will not give up.
Rewrite the thesis afterwards. That rewritten sentence is the deliverable.

Example exchange

You

My thesis is that the expansion of executive power in the twentieth century was driven primarily by war.

Primarily. So if we removed the wars, executive power stays roughly where it was in 1900?

You

I think so, the big expansions cluster around the wars.

The New Deal expansion was peacetime. Do you absorb that by changing “war” to “crisis,” or do you revise the thesis?

You

That’s fair. “Crisis” is weaker. I should keep the war claim narrower and deal with the New Deal directly.

Copyable prompt

I'm defending this thesis: [claim]. Act as a sceptical examiner in my field.
Rules: one objection at a time; be specific; track whether I quietly change the claim;
when I soften it, name the change. Do not teach me the topic, examine the thesis.

The Tell

If you finish with the same thesis you started with, unchanged, either the questioning was too soft or you never risked a real claim.

If the AI lets you redefine terms every turn without noting it, restart with a tracking rule.

The method worked when your rewritten thesis is narrower, clearer, or better evidenced.

Principle evidence

Strength of the underlying learning idea, not a claim about AI products.

The underlying learning idea is rated moderate. Sustained thesis defence is oral-examination and argumentation practice: defending a claim under follow-up questions. Evidence is pedagogical and assessment-based more than trial-dense for AI settings.

AI delivery evidence

Whether an AI tutor delivers this method well is a separate question.

Whether an AI examiner tracks drift and asks fair, field-specific questions well is a separate delivery question. Treat AI delivery as speculative here.