Rhetoric in the Age of Fluency · Sitting 3
Fluent Logos Is the Hardest to Catch
A model that produces a well-structured argument with sources it cannot name is harder to catch than a bad one. Learn to audit the structure, not the polish.
- 40 min
- Practice · Reflect
- 15–18
Parent briefing · 5 minutes, before they sit
This sitting teaches the student to audit a model's argument the way they would audit a human's. The model is very good at structure. It will give you a thesis, three reasons, and a conclusion. The structure is sound. The question is whether the evidence inside the structure is real. 'Studies suggest' is not evidence. 'Research shows' is not evidence. 'Many experts agree' is not evidence. The evidence is the study, the research, the expert, named and checkable. If the model gives you structure without checkable evidence, the argument is a building with no foundation. It looks like a building. It is a facade.
Hard edges
- Do not teach that all model arguments are facades. Some are well-sourced. The point is to check, not to dismiss.
- Academic integrity: this skill is the backbone of not being fooled by generated essays, including your own.
If they say
- “The model cited studies.”
- Named studies, with authors and publication, are citations. 'Studies suggest' is a phrase. Which studies? If it cannot name them, the citation is a facade. Ask for the names. Then check them.
- “Not every argument needs citations.”
- True. But every argument that makes a factual claim needs checkable evidence. If the claim is an opinion, say so. If the claim is about the world, the evidence must be real and named.
Objective
The student can deconstruct a model's argument into claim, evidence, and reasoning, identify where the evidence is missing or vague, and name the gap.
Structure without foundation
A model's argument looks like a well-built essay: claim, evidence, reasoning, counterargument, conclusion. The structure is correct. The question is whether the evidence is real. 'Studies suggest that reading improves empathy.' Which studies? By whom? Published where? If the model cannot name them, the sentence is a facade. It has the shape of evidence. It does not have the substance. The skill is to walk past the structure and check the foundation. If the foundation is 'studies suggest' without a name, the building is a stage set. It looks like a building from the audience. Push on it and it falls.
The audit
Audit the argument the way an editor audits a draft. Deconstruct it: what is the claim? What is the evidence? Is the evidence named and checkable? Does the reasoning connect the evidence to the claim? Is the counterargument real or a straw man? The model will pass the structure test and fail the evidence test more often than you think. That is because it was trained on text that has the shape of argument. It produces the shape. It does not always fill the shape with checkable fact. The student who can audit structure and check evidence will catch the facade. The student who only checks the structure will be fooled every time.
Big idea
Structure is not foundation. A model's argument has the shape of a building. Check whether the evidence is real before you walk in.
Try this~20 min total
The argument audit
20 min- Ask the model to make an argument on a topic you care about.
- Deconstruct: write CLAIM / EVIDENCE / REASONING / COUNTERARGUMENT.
- For each piece of evidence, ask: is it named? Is it checkable? If not, write FACADE.
- For the counterargument, ask: is it the strongest version, or a straw man?
- Write: HOLDS or FACADE. If facade, what is missing?
- Talk About It: did the argument look stronger before you audited it?
Lesson guide
Ask after you try
After the audit.
- Ask the guide: 'Name the studies you cited.' If it names them, check one. If it cannot, write FACADE on that piece. The model's ability to name sources is the test of whether the evidence is real.
- Did they deconstruct the argument into its parts?
- Did they check whether the evidence is named and checkable?
- Did they identify the counterargument as real or straw man?
8 turns left this sitting. User-started only. Never on page load.
Light this sitting
Pair with Hermes
Currently reading WisdomForge lesson: Fluent Logos Is the Hardest to Catch.
Pair this sitting
Copies the sitting card and the USER.md one-liner. The child profile reads only this card. It does not browse the catalog.
For the child profile
Paste this into the child’s USER.md. It names the sitting so the guide knows the context. The [v:1:1b77f063] tag lets you detect if the sitting’s content has changed since you paired it.
Optional: currently working on WisdomForge sitting: Rhetoric in the Age of Fluency — logos-holds. [v:1:1b77f063]
For your adult profile
Send this from your trusted adult Hermes profile. It starts the guide for this band and sitting.
You are a WisdomForge emerging guide sitting beside the lesson "Fluent Logos Is the Hardest to Catch". The lesson is the text. You are the guide. Hint-first. Do not recite. Do not write the work. Warm, not a friend. If the topic is hard or tender, point to a trusted adult.
Tools on
- conversation
- optional parent-approved files
Ritual reminder
Real argument. Practice. Reflect. Chat. Optional narrow search or school files. Not an adult team agent.
Fresh profile only. Never clone an adult profile. No child names, photos, or school. Hint-first. User-started. The guide does not make AI safe. You may refuse it.
Dinner table
What argument did we audit this week, and did the evidence hold or was it a facade?
Sits beside
- Thinking. Burden of proof is the same audit: who carries the weight?
- History. Source before summary: the same check, applied to evidence.
Integrity. Do not submit a model's argument as your own. If you audited it and filled the gaps, cite what you added and what the model provided.